Post on 17-Jun-2018
Space Weather – Impacts, Mitigation and Forecasting
Dr. Sten Odenwald
National Institute for Aerospace
Introduction
Normal, terrestrial weather is a localized phenomenon that plays out within a volume of 4
billion cubic kilometers over scales from meters to thousands of kilometers, and times as
diverse as seconds to days. Whether you use the most humble technology found in primitive
villages in Bangladesh, or the most sophisticated computer technology deployed in Downtown
Manhattan, terrestrial weather can and does have dramatic impacts all across the human
spectrum. During 2011 alone, annual severe weather events cost humanity 2000 lives and
upwards of $37 billion dollars (Berkowitz, 2011). The public reaction to terrestrial weather is
intense, and visceral, with armies of meteorologists reporting daily disturbances around the
globe, and weather forecasting models that are decades-old and have improved in reliability.
In contrast to terrestrial weather and our methods of mitigating its impact, we have the
arena of space weather, which occurs within a volume spanned by our entire solar system, over
time scales from seconds to weeks and spatial scales from meters to billions of kilometers.
Unlike the impacts caused by terrestrial weather, space weather events on the human scale are
often much more subtle, and change with the particular technology being used. There are, for
example, no known space weather events in the public literature that have directly led to the
loss of human life. The public reaction to space weather events when announced, seldom if
ever reaches the level of urgency of even an approaching, severe thunderstorm. Despite the
fact that, since the 1990s, we have become more sophisticated about communicating to the
public about the potential impacts of severe space weather, these alerts are still only consumed
and taken seriously by a very narrow segment of the population with technology at risk; satellite
owners, power grid operators, airline pilots and the like. The historical record shows that in
virtually all instances, space weather events have only led to nuisance impacts; disrupted radio
communication; occasional short-term blackouts; and occasional satellite losses that were
quickly replaced. Yet, when translated into the 21st Century, these same impacts would have a
significantly larger impact in terms of the numbers of people affected. For instance, the Galaxy 4
satellite outage in 1998 deactivated 40 million pagers in North America for several hours.
Pagers at that time were heavily used by Doctors and patients for emergency surgeries, to
name only one type of direct impact. Numerically, and in terms of function, we are substantially
less tolerant of ‘outages’ today than at any time in the past history of space weather impacts.
In this chapter, I will review the various technologies and systems that have historically
proven susceptibilities to space weather, why they are susceptible, methods being used to
mitigate these risks, and how one might estimate their social impacts. I hope to demonstrate
that, although we have a firm understanding of why technologies are at risk from basic physics
considerations, we are still a long ways from making the case that extraordinary means need to
be exerted to improve our current levels of reliability. One of the reasons for this is that we have
been living through a relatively moderate period of solar activity spanning the majority of the
Space Age. Without a major ‘Hurricane Katrina’ event in space weather, perhaps akin to the
1859 Superstorm, there is not much public outcry, commercial foresight, or political will, to
significantly improve the current preparedness situation. Moreover, the progress of technology
has been so rapid since the beginning of the Space Age in the late 1950s, that many of the
technologies that were most susceptible to space weather, such as telegraphy, compass
navigation, and short-wave communication, have largely vanished in the 21st Century, to be
replaced by substantially more secure consumer technologies.
1.0 Open-air Radio Communication
Although telegraphic communication was the dominant victim of solar geomagnetic
activity during the 1800s, by the mid-20th Century, virtually all telegraphic systems had been
replaced by land-lines carrying telephonic communications, or by the rapid rise of short-wave
broadcasting and submarine cables for trans-continental communication (Odenwald, 2010). At
its peak in ca 1989, over 130 million weekly listeners tuned-in to the BBC’s World Service. Once
the Cold War ended, short-wave broadcasting and listening went into decline. As Figure 1
shows, less than 1/3 of the stations on the air in 1970s are still operating. Compared to other
forms of communication (e.g. web-based programming) shortwave is very expensive in terms of
setting up a radio station, or providing operating costs to purchase megawatts of broadcasting
power. [Careless, 2010, 2011]. Nevertheless, by December 2011 an estimated 33% of the
human population had access to the Internet, and its vast network of formal and informal ‘news’
aggregators, including online versions of nearly all of the former shortwave broadcasting
stations.
Although shortwave broadcasting is a ghost of its former self, there are still a number of
functions that it continues to serve in the 21st Century. It is a back-up medium for ship to shore
radio, delivering state-supported propaganda to remote audiences, time signals (WWV),
encrypted diplomatic messaging, rebel-controlled, clandestine stations, and the mysterious
‘Numbers Stations’. There also continues to be a die-hard population of amateur radio ‘hams’
who continue to thrill at DXing a dwindling number of remote, low-power stations around the
world when the ionospheric conditions are optimal. Sometimes, these Ham operators serve as
the only communication resource for emergency operations. For example, during Hurricane
Katrina in 2005, over 700 ham operators formed networks with local emergency services, and
were the only medium for rapidly communicating life-saving messages. Despite the lack of
public interest or awareness of the modern shortwave band, its disruption could leave many
critical emergency services completely blind and unresponsive in a crisis.
Insert Figure 1 about here
Short wave broadcasting played such a key societal role during the first-half of the 20th
century that millions of people were intimately familiar with its quality, program scheduling,
and disruptions to this medium. Any disruption was carried as a Front Page story in even the
most prestigious newspapers such as the New York Times. Although shortwave stations were
routinely jammed by the then Soviet Union or Germany during World War II, these efforts
paled in comparison to the havoc wreaked by even a minor solar storm. Known as the
Dellinger Effect, a solar flare increases the ionization in the D and F Regions of the
ionosphere on the dayside of Earth, spanning the full sun-facing hemisphere. This absorbs
shortwave radiation but causes very low frequency (VLF) waves to be reflected. During the
four solar cycles that spanned the ‘Short Wave Era’ from 1920 to 1960, there were dozens of
flares that delivered radio blackouts, which regularly interfered with trans-Atlantic
communication; a major news and espionage flyway for information between Europe and
North America:
July 8, 1941 - Shortwave channels to Europe are affected [New York Times, p. 10]
September 19, 1941 - Major baseball game disrupted [New York Times, p. 25].
February 21, 1950 - Sun storm disrupts radio cable service [New York Times, p. 5
August 20, 1950 - Radio messages about Korean War interrupted. [New York Times, p. 5]
April 18, 1957 - World radio signals fade [New York Times, p. 25]
February 11, 1958 - Radio blackout cuts US off from rest of world. [New York Times, p. 62]
Although as we noted before, contemporary public contact with shortwave radio is nearly
zero, today there are some places where SW is still in limited use, and where the public in those
regions would be as conversant with SW fade-outs as the western world was in ca 1940. For
instance, China is expanding its SW broadcasting to remote populations across China who do
not as yet have access to other forms of communications networks. Even today, short wave
outages still make the news:
On August 9, 2011 a major solar flare caused fade-outs in the SW broadcasts of Radio
Netherlands World, but after an hour, broadcasting had returned to its normal clarity. Solar flare
disrupts RNW short wave reception [RNP, 2011]. This was the first major SW blackout in China
since the X7.9-class flare on January 21, 2005, which affected Beijing and surrounding eastern
population centers. [Xinhuanet, 2005]. On February 15, 2011 another large solar flare disrupted
southern Chinese SW broadcasting. The China Meteorological Administration reported an X2.2-
class flare at that time. [Xihuanet, 2011]. The January 23, 2012 M9-class solar flare disrupted
broadcasts on the 6 – 20 meters bands across North America, and severely affected the UHF
and VHF bands for a period of a few hours. [SWA, 2012]
1.1 Submarine Telecommunications Cables
The first copper-insulated, trans-Atlantic cable was deposited on the ocean floor in 1856
between Ireland and New Foundland, but because it was run at voltages that were too high, the
insulation broke down and the cable failed within a few weeks. The first successful cable was
laid in 1865 between Brest, France and Duxbury, Massachusetts and worked successfully for
many years, passing telegraphic signals at a speed of 2 words per minute ( 0.01 bps!). The first
copper-insulated, trans-Atlantic telephone cable was laid in 1956. By 1950, over 750,000 miles
of copper-based undersea cable had been installed between all of the major continents (ICPC,
2009]. This was followed by the first fiber optic cable TAT-8 installed between Europe and
North America in 1988. By 2009, some 500,000 miles of fiber optic cable has been deployed,
and has largely replaced all copper cable traffic due to the much higher bandwidths approaching
several terabytes/sec (See Figure 2).
Insert Figure 2 about here
Because signals degrade in strength as they travel through thousands of miles of
copper, devices called repeaters are added to the cable every 50 miles or so, and are powered
by a current flowing through a separate HV power line that parallels the cable from end to end.
Loss of power to a cable can cause immediate loss of signal, so all cables must be continuously
powered through connection to the domestic power grid or back-up generators. These voltages
can exceed 500 kV, and pose an electrocution hazard to fishing boats that accidentally snag
them. Cables are typically broken through fishing accidents, earthquakes and mechanical failure
about 150 times a year, causing a loss of communication capacity that may last from days to
weeks depending on the depth of the required repair (ICPC, 2009). Because the repair site may
only be a meter or so in length, modern repair ships routinely use GPS to reach the proper
location of the identified failed repeater, or cable damage. Also, GPS systems are used in
deploying fiber optic cables along exact, preplanned routes that minimize cable waste.
There is no formal requirement for communications companies to log cable outage
events, especially in a public archive. Consequently, outages only become public knowledge
when they impact public telecommunications activities. For example, on February 25, 2012 the
East African Marine Systems (TEAMS) data cable linking East Africa to the Middle East and
Europe was severed off the coast of Kenya by a ship that illegally dropped anchor in a restricted
area. This cable was already taking the traffic from three other fiber optic cables that had been
damaged only 10 days before. It would take three weeks before this cable could be repaired,
and data and e-commerce traffic restored to Kenya, Uganda, Rwanda, Burundi, Tanzania and
Ethiopia. [Parnell, 2012].
Copper-based submarine cables are deployed in a manner similar to the old-style
telegraph cables, and for this reason are subject to the same space weather impacts, though for
different reasons, and perhaps not the ones you might initially consider. The original telegraphic
systems and submarine cables of the 1800s were single conductors through which one-half of
the battery was connected. The other half of the battery was grounded to the local Earth to
complete the circuit! This works well when the naturally-occurring terrestrial ground current is
stable in time, and over large geographic distances comparable to the telegraph network,
however both of these conditions are badly violated during a geomagnetic storm.
During a geomagnetic storm, a strong ionospheric current appears, called the electrojet.
This current generates a secondary magnetic field that penetrates the local ground and induces
additional ground currents to flow called a Geomagnetically-Induced Current or GIC. Any single-
wire telegraph system will immediately detect this GIC, which can be much greater than the
original battery current, hence the frequent reports about mysterious high voltages and
equipment burn out. The older trans-Atlantic cables were not immune from this because they,
too, were patterned after the single-wire telegraph system and so GICs were a corresponding
problem on these systems. For example, the geomagnetic storm that occurred on 2 August
1972 produced a voltage surge of 60 volts on AT&T’s coaxial telephone cables between
Chicago and Nebraska. The magnetic disturbance had a peak rate of change of 2200 nT/min,
observed at the Geological Survey of Canada's Meanook Magnetic Observatory, near
Edmonton, and a rate of change of the magnetic field at the cable location estimated at 700
nT/min. The induced electric field at the cable was calculated to have been 7.4 V/km, exceeding
the 6.5 V/km threshold at which the line would experience a high current shutdown. [Space
Weather Canada, 2011]
One might think that modern-day fiber optic cables are immune from this GIC effect
because they involve a non-conductive optical fiber. High voltage power is supplied to the cable
at each end, with one end being at V+ and the other at V- potential. Just as for telegraph
systems, one side of the HV supply is grounded to Earth, which provides a pathway for GICs.
Repeaters for boosting the signal are connected in series along the cable axis and supplied by a
coaxial power cable. GIC currents can temporarily overload the local power supply, causing
repeaters to temporarily fail, and usually require resetting.
Have any incidents involving fiber optic cables ever been reported? We are mindful of
the old adage that absence of evidence is not the same as evidence of absence. The fact that
there is no impartial way to track outages on modern fiber optic telecommunications cables, and
there are no federal regulations that require this reporting, means that reports are voluntary.
When we search through public documents and find no cases of space weather-related cable
outages, it only means that we cannot choose between two equally likely situations: Either they
do occur and are not reported to save embarrassment, or the public records are unbiased and
so lack of examples indicated lack of an impact. There are, however, some notable examples:
“At the time of the March 1989 storm, a new transatlantic telecommunications fiber-optic cable
was in use. It did not experience a disruption, but large induced voltages were observed on the
power supply cables”. [Space Weather Canada, 2011]
1.2 Ground-based computer systems
Solar storms can be a rich source of energetic particles via shock-produced Solar Proton
Events (SPEs), galactic cosmic ray (GCR) enhancements during sunspot minimum, or events
taking place within the magnetosphere during the violent magnetic reconnection events
attending a geomagnetic storm. Although high-energy cosmic rays can penetrate to the ground
and provide about 10% of our natural radiation background, secondary neutrons can be
generated in air showers and penetrate at much higher fluxes to the ground. A number of
monitoring stations, such as the Delaware Neutron Monitor, provide day-to-day measurements
of the GCR secondary neutron background and detect ground-level enhancements (GLEs). At
aviation altitudes, these high-energy neutrons can produce avionics upsets, which are easily
corrected by error detection and correction (EDAC) algorithms or multiply-redundant avionics
systems. On the ground, and ostensibly shielded by a thick atmosphere, computer systems and
chip manufacturing processes have been allegedly affected by solar storm events (Tribble
,2010). Trying to identify even one case where such ‘computer glitches’ were caused by GCR or
space weather events remains problematical. Nevertheless, consumers and governments
expect their computer systems to function reliably (computer virus attacks excepted), so even
manufacturers such as Intel take this issue seriously. US patent 7,309,866, was assigned to
Intel for their invention of "Cosmic ray detectors for integrated circuit chips". [Hannah, 2004]
“Cosmic particles in the form of neutrons or protons can collide randomly with silicon nuclei in
the chip and fragment some of them, producing alpha-particles and other secondary particles,
including the recoiling nucleus…. Cosmic ray induced computer crashes have occurred and are
expected to increase with frequency as devices (for example, transistors) decrease in size in
chips. This problem is projected to become a major limiter of computer reliability in the next
decade.”
Bit-flip errors, in which the contents of a memory cell become switched from a ‘0’ state to
a ‘1’ state or vice versa, are a pernicious form of Single Event Upset (SEU) that continues to
plague ground based computer systems that use high-density ‘VLSI’ memory. The mechanism
is that a high-energy neutron collides with a substrate or gate nucleus, producing a burst of
secondary charged particles. These electrons and ions drift into a memory cell and increase the
stored charge until a state threshold is achieved, at which point the cell indicates a high-Q state
of ‘1’ rather than a relatively empty, low-Q state of ‘0’; hence the bit-flip error. Extensive testing
and research to identify the origin of these soft-memory errors led to alpha particle emission
from naturally occurring radioisotopes in the solder and substrate materials themselves.
Extensive re-tooling of the fabrication techniques, however, failed to completely eliminate SEUs.
Currently, a system with 1 Gby of RAM can expect one soft-memory error every week, and a 1
terabyte system can experience SEUs every few minutes. Error detection and correction
(EDAC) algorithms cost power and speed, and do not handle multi-bit errors where the parity
does not change. [Tezzaron, 2003]. According to Paul Dodd, manager for the radiation effects
branch at Sandia National Labs, "It could be happening on everyone's PC, but instead
everyone curses Microsoft. Software bugs probably cause a lot of those blue-screen problems,
but you can trace some of them back to radiation effects." [Santarini, 2005].
Although there are no specific, documented examples of ground-based computer
crashes due to specific solar storms, it is legitimate to consider what might be the societal
consequences of space weather-induced computer glitches. If they occur from time to time, it is
instructive to consider the impact that other more prosaic glitches have produced:
March 2, 2012 - Computer glitch hits Brazil’s biggest airline. Brazil’s biggest airline says a
computer glitch took down its check-in system in several airports across the country, causing
long delays [boston.com, 2012]
November 5, 2011 - HSBC systems crash affects millions across UK. HSBC was today hit by a
nationwide systems crash thought to have affected millions of customers. The bank's cash
machines, branches, debit cards, and internet banking services all stopped working at 2.45pm
after a computer glitch. [Paxman, 2011]
1.3 Space-based Computers
The first documented space weather event on a satellite occurred on Telstar-1 launched
in July1963. By November, it had suddenly ceased to operate. By exposing the ground-based
duplicate Telstar to various radiation backgrounds, Bell Telephone Laboratory engineers were
able to trace the problem to the gate of a single transistor in the satellite’s command decoder.
Apparently, excess charge had built up on the gate, and by simply turning the satellite off for a
few seconds, the problem disappeared. By January, 1963 the satellite was back in commercial
operation relaying trans-Atlantic television programs between Europe and North America (Reid,
1963).
During the 1960’s, a number of NASA reports carefully documented the scope and
nature of space weather-induced satellite and spacecraft malfunctions. There was as yet no
significant commercial investment in space, so NASA was free to analyze glitches to its own
satellites and interplanetary spacecraft. Of course military satellites of ever increasing
complexity, cost, and political sensitivity were also deployed, but no unclassified documents
were then, or are now, available to compare space weather impacts across many different
satellite platforms. This leads to an important issue that is crucial to impact assessment and
mitigation. How can we assess risks and prospective economic losses when so much of the
required data is protected through national secrecy regulations and commercial confidentiality?
Even among the ‘public domain’ NASA satellites, data as to the number and severity of ‘glitches’
is usually buried in the ‘housekeeping’ data and rarely makes it out of the daily briefing room
since it is irrelevant to the scientific data-gathering enterprise.
In a perfect world, we would like to have data for all of the 2000+ currently operational
satellites that describes the numbers, dates and types of spacecraft anomalies that they
experienced. From this we would be able to deduce how to mitigate the remaining radiation
effects, identify especially sensitive satellites and quantify their reliability, and to develop
accurate models for forecasting when specific satellites will be most vulnerable. In reality, much
of what we can learn is by ‘reading between the lines’ in news reports, correlating these biased
forms of information against the known space weather events, and hoping that a deterministic
pattern emerges. Even this has been a daunting challenge when adjacent satellites in orbit can
experience the same space weather conditions, but have very different anomalies, thereby
making correlations between space weather conditions and satellite anomalies seem less
certain.
1.3.1 How does it happen?
Satellite anomalies can be broadly defined to include any event in which some operating
mode of a satellite differs from an expected or planned condition. In this context, the term
‘anomaly’ is extremely broad, spanning a continuum of severities from trivial satellite state
changes and inconsequential data corruption, to fatal conditions leading to satellite loss. Actual
data from satellite-born sensors shows that these events can be quite numerous. For instance,
SOHO data from a 2 GBy onboard Solid State Recorder typically records > 1000 SEUs/day
(Brecca et al, 2004). Only rarely, however, do SEUs actually lead to satellite conditions requiring
operator attention – a condition commonly termed an anomaly. For SOHO, only ~60 anomalies
during an 8-year period (~ 8 anomalies/satellite/year) have required significant operator
intervention, despite the literally millions of SEU events recorded during this time.
Insert Figure 3 about here
Anomalies need not be fatal to be economically problematical. On January 20, 1994 the
Anik E1 and E2 satellites were severely affected by electrostatic discharges (ESDs). Although
the satellites were not fatally damaged, they required up to $70 million in repair costs and lost
revenue, and accrued $30 million for additional operating costs over their remaining life spans
(Bedingfield, Leach and Alexander, 1996). The Anik satellite problems were apparently the
result of a single ESD event affecting each satellite (Stassinopoulos et al., 1996), suggesting
that large numbers of anomalies are not required to 'take out' a satellite. If anomalies are
frequent enough, however, the odds of a satellite failure must also increase, as will the work
load to satellite operations. According to FUTRON (2003), satellite operators ordinarily spend
up to 40 percent of their time on anomaly-related activities. Ferris (2001) has estimated the cost
of dealing with satellite anomalies as $4,300/event leading to overall operations impacts
approaching $1 million/satellite/year under apparently routine space weather conditions.
Anecdotal reports suggest that during major solar storms, far higher operator activity can occur.
For example, the GOES-7 satellite experienced 36 anomalies on October 20, 1989 during a
single, severe solar storm event (Wilkinson, 1994).
The issue of ‘how bad can it get?’ is an interesting one, especially given our dramatically
increased reliance upon GEO satellite systems since ca 1980 that are economically baselined
on the assumption of 100% reliability during a 10 to 15 year satellite service life span. The ~250
GEO satellites now in operation produce an annual revenue of $80 billion (Ferster, 2005) so any
space weather impact is potentially costly, and can involve more than one satellite at a time.
Satellite designers use sophisticated tools to assess radiation hazards under ‘worst case’
conditions (e.g. the August 1972 and March ,1991 events) however, recent studies of extreme
space weather conditions suggest that the period since ca 1960 has not been typical of the
historical record of severe storms during the last 500 years (McCracken et al., 2001; Townsend,
2003). Moreover, there is a large discrepancy between models that predict, for example, SEU
events and actual satellite observations of them (e.g. Hoyos, Evans and Daly, 2004). Some
recent studies have attempted to estimate the economic consequences to commercial GEO
satellites for severe solar storm episodes (e.g. Odenwald and Green, 2007), but the studies
were hampered by the lack of detailed knowledge of how the frequencies of satellite anomalies
vary in severity with storm intensity. Consequently, the loss of a satellite during a severe space
weather event could not be modeled realistically, nor its economic impact properly assessed.
Most reported anomalies, broadly defined, are nuisances involving recoverable data
corruption, easily-corrected phantom commands, or 'bit flips' often caught by onboard EDAC
algorithms. These are not the kinds of anomalies that lead to significant economic
consequences for a commercial satellite. Other less frequent anomalies cause sub-system
failures, out-of-spec satellite operations, attitude and telemetry-lock errors and even outright
satellite failures. These are most certainly the kinds of anomalies that have economic
consequences. Some authors have also classified anomalies by satellite orbital location (e.g.
LEO, MEO, GEO), recognizing that each environment has its own physical drivers for anomaly
generation, but more often than not, these classes are aggregated together. Here is one
possible scheme:
Class 1 - Mission-Failure - The satellite ceases operation as a consequence of an
unrecoverable system malfunction. ( e.g. Telstar-401)
Class 2 - Mission interruption - involves a recoverable damage to sub-systems. Only built-in
redundancies, if available, are capable of mitigating some of these problems, where the
satellite's safe mode may be enabled, or a back-up subsystem has to be activated (e.g. Anik-
E1). These may take hours of effort to remedy, at a cost to satellite revenue and operator
overhead charges.
Class 3 - Performance decrease - can include spacecraft pointing errors, attitude control
system error, or a brief loss of data or telemetry usually corrected by a manual or automatic
system reset.
Class 4 - Inconsequential - memory bit-flips and switching errors easily corrected using on-
board EDAC software, or simple operator action. (e.g. TDRSS-1 telemetry; cosmic ray
corruption of Hubble Space Telescope data).
One of the earliest, and most detailed, publically available studies of satellite anomalies
and reliability is the work by Hecht and Hecht (1986 : the Hecht Report). The study was based
on 2,593 anomaly reports for 300 satellites launched between ca 1960 and 1984. There were
~350 satellites in operation by ca 1984, so the Hecht Report is relatively complete. This ground-
breaking study analyzed the detailed reports provided by 96 satellite programs. A 'failure' was
defined as "…the loss of operation of any function, part, component or subsystem, whether or
not redundancy allowed the recovery of operation”. Their study identified 213 Class 1 and 192
Class 2 anomalies out of a total collection of 2593 anomalies for a mission failure rate defined
by our Class 1 of about 405/2593 or 1 in 6. No attempt was made to correlate the anomalies
with space weather conditions.
One of the most widely used, recent starting points for anomaly studies is the archive
assembled by Wilkinson and Allen (1997; hereafter NGDC) which identifies most of the 259
satellites by name, or code, along with orbital location and/or altitude information. The date and
type of anomaly is provided for many of the 5,033 events spanning the time period from 1970 to
1997, so that a proper assessment can be attempted of the various category-specific anomaly
rates as a function of date and satellite type. There are 3,640 events that have been tagged
according to type and system impact, including 647 SEU events and 848 ESD events. The
NGDC archive contains 43 commercial GEO satellites included in the archive, accounting for a
total of 480 anomalies spanning 20 years, and also appears to contain about 40% of all
operating satellites during the sample time span, and is relatively complete for our purposes.
The average annual anomaly rate of the GEO satellites was found to be about 3
anomalies/satellite/year, but can rise to twice or three times this rate during enhanced space
weather conditions.
Robertson and Stoneking (2005: Goddard) examined 128 severe (Class 1 and 2)
anomalies among 764 satellites. The data were culled from web-based satellite anomaly lists
including the 'Airclaims Space Track' as well as NASA documents and the Aerospace
Corporation 'Space Systems Engineering Database', and only included satellites from the US,
Europe, Japan or Canada. The total number of satellites (military + commercial) operating
during this interval is 827, so the sample contains about 92 % of all possible operational
systems during the 1990-2001 time period. A total of 35 anomalies were Class 1, which led to
what was considered the total loss of the satellites. About one anomaly in four (Class 1 vs Class
1 + 2) is of the severe Class 1. Their calculated anomaly rate was based on the number of
anomalies recorded, divided by the number of satellites launched during a given year. Re-
normalizing their mishap rates to, instead, reflect the annual operating satellites, the average
mishap rate for Classes 1+2 is about 0.019 +/- 0.006 anomalies/sat/year. The inverse of this
rate is 166 which is sometimes called the mean time to failure (MTF). Clearly for commercial
satellites expected to last 10 to 15 years before replacement, a MTF of 166 years is good news!
The correlation between these anomalies and space weather events was not studied.
The extensive studies by Belov et al (2004), and Dorman et al (2004), included satellite
anomaly reports based on 300 satellites and ~6,000 anomalies spanning the time period from
1971 to 1994. The data was drawn from NASA archives, the NGDS archive and unpublished
reports from 49 Kosmos satellites (1971-1997). The term 'anomaly' was never precisely defined,
but since the survey included the NGDC archive without distinction, we can assume that all
Class 1-4 events were grouped together. The sample included 136 satellites in GEO orbits.
They deduced that there were typically 1 to 10 anomalies/satellite/year. Specifically, the LEO
Kosmos satellites experienced 1-7 anomalies/satellite/year, however some Kosmos satellites
(Kosmos 1992 and 2056) reported ~30 anomalies/satellite/year. Their statistical analysis
indicated that anomalies occur during days when specific space weather parameters
(electron/proton fluxes, Dst, Ap, etc) are disturbed. The largest increases coincide with times
when electron and proton fluences are large, and can cause up to 50-times enhancements in
anomalies over quiet-time conditions. There appears to be a threshold of 1,000 pfu (E > 10
Mev) for proton fluxes, below which there are few anomalies reported. The anomalies continue
to remain high for two days after the SPE event.
Koons et al. (1999) published 'The Impact of the Space Environment on Space
Systems', which investigated a sample of 326 anomaly 'records' collected from a diverse
assortment of satellites culled from the NGDC 'Satellite Anomaly Manager', Orbital Data
Acquisition Program (ODAP: Aerospace Corp), NASA's Anomaly reports (Bedingfield et al.
1996, and Leach and Alexander, 1997), and the USAF Anomaly Database maintained by the
55th Space Weather Squadron. The specific number of satellites involved was not stated,
however the ODAP archive contains information from 15 USAF and 91 non-Air Force 'programs'
no doubt drawn from LEO, MEO and GEO satellite populations. Although no information was
provided as to the time period spanned by the study, the individual archives extend from 1970 to
1997. The definition of a Record in terms of anomalies can vary enormously. Each record
contained information for one class of anomalies for one 'vehicle'. Anomalies of a similar class
were of the same functional type. Approximately 299 records out of 326 (92%) have causes
diagnosed as 'space environment' but this does not necessarily correlate with a count based on
anomaly frequencies. An example cited is that one record for the MARECS-A satellite included
617 anomalies. About 51/326 records were from commercial satellite systems and programs. In
terms of the distribution of the records with anomaly diagnosis, 162 (= 49%) were associated
with Electrostatic Discharges, 85 (= 26%) with SEUs, and 16 ( =5%) with 'total radiation
damage'. Based on 173 reports of how quickly the anomalies were rectified, the Koons et al
(1999) study indicates that the number of mission failures represents 9/173 reports for a
frequency rate of 1 in 19. The rates for the other classes are: Class 2 (More than 1 week) =
39%; Class 3 (1 hr to 1 week) = 35% and Class 4 (Less than 1 hour) = 20%.
Ferris (2001) analyzed 9,200 satellite operations discrepancy reports from 11 satellites
between 1992-2001. A 'discrepancy' was defined as "...the perception by a satellite operator
that some portion of the space system had failed to operate as desired." The satellites were
selected on the basis of which operators and owners were willing to divulge detailed anomaly
logs for this study, which is a strong bias probably in favor of systems that had low absolute
rates and few critical failures. Only three of the satellites were communications satellites; none
were for civilian commercial use. This, of itself, is a problem since we cannot know to what
extent these satellites are typical, or whether they are pathological. This is often the case when
working with studies in which the satellite identities are not publically revealed. Of the
discrepancies catalogued, only 13% involved the satellites themselves. The vast majority, 48%,
involved issues with the ground segment, and specifically, most were discrepancies generated
by software issues (~61% of total discrepancies). Typical discrepancy rates involving 1,200
events imply ~13 discrepancies/satellite/year. There were, however, higher rates recorded in
1996 involving 160 events for 4 satellites for a rate of 40 discrepancies/satellite/year or about
one every 9 days. The study was the first one published in the open literature that also provided
an assessment of the cost of rectifying these anomalies. Routine problems that require no more
than 10 minutes to resolve by a team of 8 people cost $800 per event. More significant
problems requiring 3-8 hours and more people cost $4,300 per event. The estimate only
included man-hours and an average of the resolution times for the logged events, and not the
cost of equipment or materials. In the latter case an 'event' may include the replacement of part
of the ground station, processors or other mechanical items.
Cho and Nozaki (2005) investigated the frequency of ESDs on the solar panels of five
LANL satellites between 1993 – 2003. During this period, LANL 1989-046 experienced 6038
ESDs/year while LANL-92A recorded 290 ESDs each year. Although the cumulative lifetime
ESD rates on solar panels can exceed 6,000 events/kW over 15 years, the chances of a
catastrophic satellite failure involving substantial loss of satellite power, remains small, though
not negligible. For example, in 1973, the DSCS-9431 satellite failed as a result of an ESD event.
More recently, the Tempo-2 (1998) and ADEOS-2 (2003) satellites were also similarly lost.
Koons et al (1991, 2000) and Dorman et al (2005) have shown that ESDs appear to be
ultimately responsible for half of all mission failures (e.g. Class 1 anomalies) and correlated with
space weather events.
Wahlund et al (1999) have studied 291 ESD events on the Freja satellite (MEO orbit)
and have found that the number of ESDs increases with increasing Kp. A similar relationship
between increasing Kp and anomaly frequency was found by Fennell et al (2000) for the
SCATHA satellite (near-GEO orbit). These results are consistent with earlier GOES-4 and 5
satellite studies by Farthing et al. (1982) and by Mullen et al. (1986). In addition to Kp, Fennell
et al (2000) and Wrenn, Rogers and Ryden (2002) identified a correlation between 300 keV
electron fluxes and the probability of internal ESDs from the SCATHA satellite. The probability
increases dramatically for electron fluxes in excess of 100,000 pfu. A similar result was found a
number of years earlier by Vampola (1987). At daily total fluences of ~1012 electrons/cm2 the
probability of an ESD occurring on a satellite exponentially reaches 100% (e.g. Baker, 2000).
1.3.2 That Was Then – This is Now.
During the 23rd Sunspot Cycle (1996-2008) there were dozens of satellite malfunctions
and failures noted soon after a major solar storm event, beginning with Telstar-401 (1996) and
ending with the Japanese research satellite ASCA on October 29, 2003. The current ‘24th’ cycle
has had its own satellite outages and malfunctions of note, with the majority of the solar activity
still several years in the future.
On August 25, 2011 South Africa's $13 million LEO satellite SumbandilaSat failed, and
the explicit cause was stated publically to be 'damage from a recent solar storm', which caused
the satellite's onboard computer to stop responding to commands from the ground station. This
was not, however, the first time this satellite was damaged by radiation. Shortly after its launch
in September 2009, radiation caused a power distribution failure that rendered the Z-axis and
Y-axis wheel permanently inoperable, meaning that the craft tumbles as it orbits and has lost
the ability to capture imagery from the green, blue and xantrophyll spectral bands. The reason
given for the lack of proper radiation hardening was that there was not enough money to do this
properly, and the satellite was built from commercial off-the-shelf (COTS) equipment.
Moreover, SumbandilaSat was intended only as a technology demonstrator. [Martin, 2012]
The case of the Anik F2 'technical anomaly' on October 6, 2011 is a replay of similar
stories during the 23rd cycle. The satellite entered a Safe Mode that caused it to stop
functioning and turn away from Earth. The Boeing satellite was launched in 2004 and was
expected to function for 15 years. The owner of the satellite, Telsat, indicated in public news
articles that they did not believe the problem had to do with the arrival of a CME that reached
Earth early the same morning, but was caused by some other unspecified internal issue with the
satellite itself. It is the first serious anomaly of its kind since the satellite was launched in 2004.
What the news reports failed to mention was that the sun has been relatively quiet for the
majority of this 7 year period. [Mack, 2011]
The temporary outage of Anik F2 caused a number of problems that impacted millions of
people covered by this satellite service. WildBlue satellite ISP in the United States uses Anik F2
to provide broadband services to about a third of its customers. A total of more than 420,000
subscribing households mostly in parts of rural America lost service for several days, along with
ATM service. Canadian Broadcasting Corporation indicated that 39 rural communities, and
7,800 people lost long-distance phone service. The satellite is also used for air traffic control,
causing the grounding of 48 First Air flights, and 1000 passengers, in northern Canada.
Communities in the North West Territories were instructed to activate their emergency response
committees, and start using their Iridium phones. [Mack, 2012; CBS News, 2012; Marowits,
2011]
April 5, 2010 - Galaxy 15 experienced an electrostatic discharge that caused a severe
malfunction, rendering the satellite capable of re-transmitting any received signal at full-power,
but not able to receive new commanding [de Shelding, 2011]. Reports cited a space weather
event on April 5 as the probable cause of the electrostatic discharge that was the likely
triggering event, however although Intelsat acknowledged the ESD origin, they categorically
refuted the space weather cause in the April 5 solar event, preferring to declare that the orgin of
the ESD was unknown. A consequence of this type of satellite failure is that Galaxy-15 was
potentially able to interfere with other GEO satellites as it came within 0.5 degrees of their
orbital slots. Thanks to careful, and complex, maneuvering of the satellites to maximize their
distance from this satellite as it entered their orbital slots, AMC-11, Galaxy-13, Galaxy-18,
Galaxy-23 and SatMex-6 and Anik F3 were able to reduce or eliminate interference, and no
impacts to broadcasting were reported or acknowledged. "The fact that you haven't heard about
channels lost or interference is the proof that we have been able to avoid issues operationally,"
said Nick Mitsis, an Intelsat spokesperson. "I don't want to underplay that." [Clark, 2010]. In
January 2011 commanding of the satellite resumed and its ’zombisat’ moniker has been
changed to ‘phoenix’.
1.4 Cellular and Satellite Telephones
Although telephone calls by land lines are among the safest communication technology,
and the most resistant to space weather effects, they have also been in rapid decline thanks to
the wide spread adoption of cellular and mobile phones, especially among the under-30
population. According to an article in The Economist [2009] customers are discontinuing
landline subscriptions at a rate of 700,000 per month, and that by 2025 this technology will have
gone the way of telegraphy. Between 2005 and 2009, the number of households with cell
phone-only subscriptions rose from 7% to 20%. In terms of space weather vulnerability, there is
one important caveat. Without an electrical power grid, conventional land-lines fail, and cell
phones may not be recharged even though the cell towers may have emergency back up power
capability. An example of this vulnerability occurs whenever natural disasters strike and cell
towers are unavailable, or the crushing load of cell traffic renders the local tower network
unusable. Moreover, one does not have to wait for power grid failure to have an impact on cell
phone access during episodes of solar activity.
Insert Figure 4 about here
A seminal paper by Lanzerotti et al. (2005) demonstrates that solar radio bursts, which
occur rather often in an active photosphere, can cause enhanced noise at the frequencies used
by cellphones (900 MHz to 1900 MHz), when the observer’s angle between the cell tower and
the sun is small. This interference effect shows up in the Dropped-Call statistic for east-facing
receivers at sunrise or west-facing receivers at sunset. For a given cell phone and cell tower in
the optimal line-of-sight geometry with respect to the sun on the horizon, dropped calls occur
about once every 3 days during solar maximum, and every 18 days during solar minimum. The
article notes that the detailed, direct, evidence for solar-burst influence on cell phones remains a
proprietary issue not openly available for investigation. The authors note that "solar bursts
exceeding about 1000 sfu (solar flux units, 1 sfu = 10−22 W m−2 Hz−1) can potentially cause
significant interference when the Sun is within the base-station antenna beam, which can
happen for east- or west-facing antennas during sunrise and sunset at certain times of the
year." Because base stations are only vulnerable for about two hours each day during sunrise
and sunset, a typical station might be affected about one day out of 42 for solar maximum, and
one day in 222 during solar minimum.
1.5 GPS-based Systems
Navigation by satellite is not a new technology, and was first introduced by the US Navy
in 1960 with the orbiting of five Transit satellites. This system was replaced by the NAVSTAR-
GPS system in the 1970s. The first commercial use of satellite-based global positioning systems
came less than 1 year after the next generation, 24-satellite 'Block I-GPS' constellation had
been deployed in 1994, when Oldsmobile offered the GuideStar navigation system for its high-
end automobiles. The GPS satellites provided an L1 channel at 1575 MHz capable of 10-meter-
scale precision, that in 1990 was 'selectively degraded' to 100-meter precision. In 1999,
President Clinton ordered that selective availability be turned off, and on May 1, 2000 the
modern era of non-military GPS was ushered-in. Since 2000, the commercial applications of
GPS have enormously expanded to include, not only car navigation aids, but oil extraction, fiber
optic cable deployment, civilian aviation, emergency services, and even expanding public
cellphone services, called apps, to locate nearby stores, restaurants and even parking spaces in
downtown Manhattan! A report by Berg Insight (2011) indicates that GPS-enabled mobile
phones reached 300 million units in 2011, and is expected to reach nearly 1 billion units by
2015.
Although the GPS constellation is stationed in polar orbits that frequently pass through
the van Allen radiation belts in MEO, they are well-shielded and are upgraded every 5-10 years
through replacement satellites such as the Block-II and Block-IIII systems. Although the details
of the frequency of satellite anomalies is highly classified, it can be surmised that a legacy of 40
years of space operations has left the GPS system with a broad assortment of mitigation
strategies for essentially eliminating outages. Nevertheless, there is one aspect of GPS system
operation that cannot be so easily eliminated.
GPS signals must be delivered to ground stations by passage through the ionosphere.
Because radio propagation through an ionized medium causes signal delays, and accurate
timing signals are important in locating a receiver in 3-dimensional space, any changes in
ionospheric electron content along the propagation path will cause position errors in the final
solution. Space weather events, especially X-ray flares, cause increased ionization and
introduce time-dependent propagation delays that can last for many hours until the excess
ionospheric charge is dissipated through recombination. This also causes amplitude and phase
variations called scintillation, which causes GPS receivers to loose lock on a satellite. Since a
minimum of 4 satellites are required to determine a position, excess scintillation can result not
just in a bad position solution, but can cause a loss-of-lock so that not enough satellites are
available for various locations at various times during the event.
When civilian, single-frequency GPS systems using the L1 frequency are used, the
anomalous propagation problem has to be mitigated by reference to a 'GPS Ionospheric
Broadcast Model' and making the appropriate corrections. The resulting accuracy is about 5
meters. But this correction can only work for a limited period of time and so the path-delay
problem is only partially solved. The result is that most civilian GPS systems can be easily
disturbed by solar activity. Dual-frequency GPS systems that operate at L1 (1575 MHz) and L2
(1228 MHz ) can measure the differential propagation of the satellite signal in real-time, and by
relating this to the plasma dispersion equation, calculate the instantaneous total electron
content (TEC) along a path, and then use this to make the requisite on-the-spot timing
correction. In fact, this method can be turned around by using networks of GPS receivers to
actually map out the changing ionospheric structure over many geographic locations. Figure 5
shows one such 'TEC' calculation for April 20, 2012 for 19:00 UT developed by JPL. The black
spots are the GPS receivers in the network. Green indicates a TEC of about 50 x 1016
electrons/meter2 while red indicates 80 x 1016 electrons/meter2. Generally, a TEC of 6 x 1016
electrons/meter2 corresponds to an uncorrected position error of about 1 meter. The figure
displays potential position errors as high as 13 meters over Chile.
Insert Figure 5 about here
Although the L1 carrier signal can be received without special instrumentation, the L2
timing information is coded and not accessible to non-military receivers. However, by using a
technique called differential GPS, civilian GPS systems now rival, or even exceed, military
precision in those areas where the requisite DGPS ground reference stations are available. If
you are navigating in a large city, DGPS is probably available to you, but if you are 'in the
middle of nowhere' chances are you only have single-frequency GPS to guide you.
We have already discussed this briefly in the context of GPS signal propagation and
ionospheric scintillation. Because many space weather phenomena couple efficiently to the
ionosphere, it is unsurprising that space weather issues have always been foremost in the
discussion of GPS accuracy and reliability even apart from the fact that the GPS satellites
themselves are frequently located in one of the most 'radio-active' regions of the
magnetosphere.
One of the first unclassified studies to quantitatively assess GPS behavior under solar
storm conditions was conducted, inadvertently, by NOAA in 2001. They had set up a network of
70 GPS receivers from Alaska to Florida to test a new weather observation and climate
monitoring system called the GPS-MET Demonstration Network. A major geomagnetic storm
between March 30 and 31 caused significant changes in the GPS formal error, and was
correlated with the published Kp index during the course of the event [NOAA, 2001]. Since
then, a variety of anomalous changes in GPS precision have been definitively traced to, and
found to be correlated with, geomagnetic storms and solar flare events. This also means that
systems that rely on GPS for high-precision positioning have almost routinely reported
operational upsets of one kind or another. For example [NOAA, 2004]:
October 29, 2003 - the FAA’s GPS-based Wide Area Augmentation System (WAAS) was
severely affected. The ionosphere was so disturbed that the vertical error limit was
exceeded, rendering WAAS unusable. The drillship GSF C.R. Luigs encountered significant
differential GPS (DGPS) interruptions because of solar activity. These interruptions made
the DGPS solutions unreliable. The drillship ended up using its acoustic array at the seabed
as the primary solution for positioning when the DGPS solutions were affected by space
weather.
December 6, 2006 - the largest solar radio burst ever recorded affected GPS receivers
over the entire sunlit side of the Earth. There was a widespread loss of GPS in the
mountain states region, specifically around the four corners region of New Mexico and
Colorado. Several aircraft reported losing lock on GPS. This event was the first of its kind to
be detected on the FAA, WAAS network.
Apart from changes in ionospheric propagation, we have the problem that, if the GPS
signal cannot be detected by the ground station, and the minimum of 4 satellites is not detected,
a position solution will not be available at any accuracy. This situation can arise if the GPS
signal is actively blocked or jammed, or if the natural background radio noise level at the L1 and
L2 frequencies is too high. This can easily happen during radio outbursts that accompany solar
flare events. This happened the day after the December 5, 2006 solar flare, and was intensively
studied by Kintner at Cornell, and presented at the Space Weather Enterprise Forum in
Washington, DC on April 4, 2007 [NOAA, 2007].
1.6 Electrical Power Grids
Although the issue of space weather impacts to the electrical power grid will be covered
more extensively in a future volume, we will review the main points of this vulnerability, provide
concrete examples, and review briefly the impacts and consequences of future large
geomagnetic storms.
It has been well known for decades that geomagnetic storms causes changes in the
terrestrial ground current. The most dramatic examples of this effect are in the many reports of
telegraph system failures during the 1800s. So long as a system requires an 'earth ground', its
circuit is vulnerable to the intrusion of geomagnetically-induced currents (GICs). For the electric
power grid, these DC currents do not need to exceed much above 100 amperes in order to do
damage (Odenwald,1999, Kappenmann, 2010 ).
When GICs enter a transformer, the added DC current causes the relationship between
the AC voltage and current to change. It only takes a hundred amperes of GIC current or less to
cause a transformer to overload during one-half of its 60-cycle operation. As the transformer
switches 120 times a second between being saturated and unsaturated, the normal hum of a
transformer becomes a raucous, crackling whine physicists call magnetostriction.
Magnetostriction generates hot spots inside the transformer where temperatures can increase
very rapidly to hundreds of degrees in only a few minutes, and last for many hours at a time.
During the March 1989 storm, a transformer at a nuclear plant in New Jersey was damaged
beyond repair as its insulation gave way after years of cumulative GIC damage. During the 1972
storm, Allegheny Power detected transformer temperature of more than 340 F (171 C). Other
transformers have reached temperatures as high as 750 F (400 C). Insulation damage is a
cumulative process over the course of many GICs, and it is easy to see how cumulative solar
storm and geomagnetic effects were overlooked in the past.
Outright transformer failures are much more frequent in geographic regions where GICs
are common. The Northeastern US with the highest rate of detected geomagnetic activity led
the pack with 60% more failures. Not only that, but the average working lifetimes of transformers
is also shorter in regions with greater geomagnetic storm activity. The rise and fall of these
transformer failures even follows a solar activity pattern of roughly 11 years.
The connection between space weather events and terrestrial electrical systems has
been documented a number of times. Some of these examples are legendary (1989, 2003)
while others are obscure (1903, 1921). Given the great number of geomagnetic storms that
have occurred during the last 100 years, and the infrequency of major power outages, this
suggests that blackouts following a major geomagnetic storm are actually quite rare events.
Consider the following historical cases:
November 1, 1903 - The first public mention that electrical power systems could be
disrupted by solar storms appeared in the New York Times, November 2, 1903 "Electric
Phenomena in Parts of Europe". The article described the, by now, usual details of how
communication channels in France were badly affected by the magnetic storm, but the article
then mentions how in Geneva Switzerland, [NYT,1903]
"...All the electrical streetcars were brought to a sudden standstill, and the unexpected
cessation of the electrical current caused consternation at the generating works where all efforts
to discover the cause were fruitless".
May 15, 1921 - The entire signal and switching system of the New York Central Railroad
below 125th street was put out of operation, followed by a fire in the control tower at 57th Street
and Park Avenue. The cause of the outage was later ascribed to a ‘ground current’ that had
invaded the electrical system. Brewster New York, railroad officials formally assigned blame for
a fire destroyed the Central New England Railroad station, to the aurora. [NYT,1921]
August 2, 1972 - The Bureau of Reclamation power station in Watertown, South Dakota
experienced 25,000-volt swings in its power lines. Similar disruptions were reported by
Wisconsin Power and Light, Madison Gas and Electric, and Wisconsin Public Service
Corporation. The calamity from this one storm didn't end in Wisconsin. In Newfoundland,
induced ground currents activated protective relays at the Bowater Power Company. A 230,000-
volt transformer at the British Columbia Hydro and Power Authority actually exploded. The
Manitoba Hydro Company recorded 120-megawatt power drops in a matter of a few minutes in
the power it was supplying to Minnesota.
March 13, 1989 - The Quebec Blackout Storm - Most newspapers that reported this
event considered the spectacular aurora to be the most newsworthy aspect of the storm. Seen
as far south as Florida and Cuba, the vast majority of people in the Northern Hemisphere had
never seen such a spectacle in recent memory. At 2:45 AM on March 13, electrical ground
currents created by the magnetic storm found their way into the power grid of the Hydro-Quebec
Power Authority. Network regulation failed within a few seconds as automatic protective
systems took them off-line one by one. The entire 9,500 megawatt output from Hydro-Quebec's
La Grande Hydroelectric Complex found itself without proper regulation. Power swings tripped
the supply lines from the 2000 megawatt Churchill Falls generation complex, and 18 seconds
later, the entire Quebec power grid collapsed. Six million people were affected as they woke to
find no electricity to see them through a cold Quebec wintry night. People were trapped in
darkened office buildings and elevators, stumbling around to find their way out. Traffic lights
stopped working, Engineers from the major North American power companies were worried too.
Some would later conclude that this could easily have been a $6 billion catastrophe affecting
most US East Coast cities. All that prevented the cascade from affecting the United States were
a few dozen capacitors on the Allegheny Network (Odenwald, 1999).
October 30, 2003 - Malmo Sweden, population 50,000 lost electrical power for 50
minutes [Pulkkinen et al., 2005]. The blackout was caused by the tripping of a 130 kV line. It
resulted from the operation of a relay that had a higher sensitivity to the third harmonic (=150
Hz) than to the fundamental frequency (=50 Hz). The excessive amount of the third harmonics
in the system has been concluded to have resulted from transformer saturation caused by GIC.
Currents as high as 330 Amperes were recorded on the Simpevarp-1 transformer. [Wik et al.
,2009]
October, 2003 - South Africa Transformer Damage. The ESKOM Network reported that
15 transformers were damaged by high GIC currents. Figure 6 shows one of the transformers in
a view reminiscent of the legendary images of the 1989 Quebec transformer failure. [Murtagh,
2009]
Insert Figure 6 about here
Extensive studies have already been conducted on the most cost-effective means for
reducing or eliminating GICs in electric power grid components [Kappenman, 2010]. The
strategies generally include adding individual capacitors to each of the transformer HV lines, or
adding a blocking resistor or capacitor to the ground lines in all transformers. Blocking
capacitors were, for example, installed on the entire Hydro-Quebec power grid following the
March 1989 blackout, as well as the WECC region in the western US. Although this strategy
seemed to be successful in reducing GICS and reactive power on some of the lines, the impact
was deemed only modest, 12 to 20% for the WECC network with 50% penetration, given the
cost expended. Adding blocking capacitors to the transformer neutral ground connector (points
A and B in the figure) is the simplest and most direct method for achieving a 100% reduction in
DC, GICs from transformer primaries, but this method is known to alter the impedance of the
network in unpredictable ways as the devices are selectively deployed rather than universally
adopted.
The next most direct, and also the most cost-effective method is by adding a low-
ohmage and low-voltage resistor to the neutral ground of each 3-phase transformer (see red
boxes in figure). Preliminary studies [Kappenman, 2010] suggest that this method could
achieve a 60% reduction in GIC amperages to transformer primaries. The cost would be at most
$100,000 per transformer in the US power grid, which contains some 5000 transformers, for a
total cost of about $500 million. A simulation of the Hydro-Quebec power grid during the 1989
failure, but with neutral ground resistors installed reveals a dramatic reduction in the GICs to
which the 45 transformers in the 735 kV grid were subjected, with hypothetical 10-ohm blocking
resistors reducing the GICs from 550 amps to only about 75 amps. A comparison of the
application of 5-ohm blocking resistors to the US power grid during the March 1989
geomagnetic storm is shown in Figure 1.
The maximum storm time disturbance was about 450 nT/minute, but even with proper
mitigation, the US grid may not be immune from the largest known geomagnetic events,
although the severity of the impact could be reduced by 60% from the case where no such
mitigation is implemented. During the 1921 storm, a disturbance field of 4800 nT/min was
estimated. Without mitigation, over 500 transformers would be damaged, but with mitigation
only about 40 would be damaged according to these simulations [Kappenman, 2010]. This ten-
fold reduction is not inconsequential.
It is also worth mentioning that, although blackouts are a dramatic consequence of
severe GICs caused by space weather, economic consequences also flow from the on-going
stresses to the power grid during non-black out conditions. For example, Forbes and St. Cyr
(2004) note that the constant impacts of minor space weather events over a long period of time
disrupts the system that transmits the power from where it is generated to where it is distributed
to customers. In examining the determinants of the real-time electricity market price over the
period June 1, 2000, through December 31, 2001, they concluded that solar storms (over this
period) increased the wholesale price of electricity by approximately 3.7 percent or
approximately $500 million. Kappenmann (2012) has recently shown that in the months
following the March 1989 Quebec event, a statistically significant number of transformers in the
United States had to be prematurely replaced, with the greater number of replacements found in
proximity to the Quebec power grid.
Of course, not all electrical power blackouts have anything to do with space weather.
Most of us have experienced at least on ‘outage’, and in some regions like Washington, DC, it is
typical to have 3-5 outages every year lasting from hours to days. Hamachi-LaCommare and
Eto (2004) have studied the economic costs of annual power outages and power ‘sags’ and
have found that they cost as much as $130 billion annually to the GDP. We are accustomed to
electrical blackouts and quietly absorb them into our economy, with some grumbling about lost
food and time. The long term trends for normal blackouts also points to the progressive failures
inherent to an ageing domestic power grid [Karn, 2007]. The over use of this resource is
highlighted by the dramatic growth in bulk power transactions on, for example, the Tennessee
Valley Authority system which exploded from less than 20,000 such transactions in 1996 to
more than 250,000 by the end of 2001. [DOE, 2005].
Increased bulk power transactions have led to a substantial drop in capacity margin,
which provides little room either for growth or to maneuver in times of crisis. By some accounts
[Patterson, 2010] there were 41 blackouts nationwide between 1991-1995, and 92 between
2001-2005. In 2011 alone, there were 109 affecting communities of 50,000 or more people. The
Eaton Corporation, an agrigator of news and industry reports of blackouts across the US states
finds that between 2009 and 2011, the number of power outages rose from 2,169 to 3,041 and
the number of people impacted climbed from 26 million to 42 million [Eaton, 2011].
A ‘typical’ person comes into contact with the following technologies each day: cell
phones, portable computing, credit card verification (ATM), navigation (GPS), electrical utilities
(water pumps, gasoline pumps, hospital facilities, home lighting, city electrification, cell phones-
recharging). All of these ‘essential’ systems rely on electricity either at the point of creation
(satellite GPS and ATM verification) or at the point of delivery (cell phone, gas pump, water,
etc). All are expected to be ready when needed with 100% reliability. In recent human history,
we have been successful in delivering these services even in the face of a number of space
weather events. The lynchpin technology is, of course the electric power grid which citizens use
to ‘tap into’ essential communication and utility resources. It is unlikely that even a Superstorm
event will dramatically impact the number of satellites operational, and backup transponders are
readily available in case of emergencies. The ubiquitous cell phone would not fail if satellites
failed, but satellites do carry the bulk of financial transactions, GPS and military CCC traffic. The
loss of key satellites, or a critical number, would render these services reduced in capability.en
made of the cascading problems involved in ‘re-booting’ such a large grid, especially in the
event of component failures and burn-outs which would necessitate replacement, not on a local
scale, but quite possibly on a global scale, with only a few key manufacturers able to service
these needs.
1.7 Airline Travel
Generally, the known routes for space weather impacts to aviation are through
passenger safety (radiation), flight avionics (computer/system glitches), communications (radio
interference) and scheduling (delays, route changes). Historically there have been anecdotal
instances of each of these being identified. For example,
July 19, 1947 - Sunspots delay planes [New York Times, July 18, 1947 p.15]
Although earlier flight navigation methods involved compass bearings and LORAN-C,
which in principle could be affected by geomagnetic storms and shortwave interference, there
are actually no known instances where space weather events caused significant disruptions to
these navigation technologies. Today, however, the airline industry is adopting GPS navigation
as the new standard, and its implementation in the Wide-Area Augmentation System (WAAS).
WAAS is a combination of GPS and local, ground based metrology reference stations that
provides 1-meter lateral and 1.5-meter vertical position resolution every 6 seconds for aircraft
flying over the continental US, Canada and Alaska. There are also several satellites, such as
Galaxy-15 that are involved in the WAAS system as part of its space-leg. Because its uses
GPS satellites, the WAAS system is not immune to space weather effects that impact the
ionosphere. Consequently during a severe storm event, WAAS may not be available for several
minutes, or even hours, in some regions of the normal coverage area [Doherty, 2011]. Studies
have shown that approaches with vertical guidance (APV) are restricted during times of
geomagnetic activity in terms of the APV coverage with Dst during the period from July 2003 to
March 2004. For airports in which APV coverage is not available, flights must revert to IFR or
VFR landing regulations within a vertical distance of 200 meters of the runway.
A number of national and international studies have been conducted to assess the
radiation load on passengers during active space weather conditions and under otherwise
normal circumstances. Normal background radiation doses are typically 0.3 microSv/hr or 3
mSv/year. For passengers and flight crews, the actual cabin exposure varies with the
geographic latitude of the flight, the altitude of the flight, and the combined GCR and solar fluxes
of particles. For example, Bottollier-Depois et al (2000) determined from direct measurements at
maximum solar activity in 1991-1992 and at minimum activity in 1996-1998. The lowest mean
dose rate measured was 3 microSv/hr during a Paris-Buenos Aires flight in 1991. The highest
rates were 6.6 microSv/hr during a Paris-Tokyo flight on a Siberian route and 9.7 microSv/hr on
Concorde in 1996-1997. A number of similar studies since then have supported the idea that
there is in fact some additional passenger and flight crew radiation exposure caused by space
weather, however the levels are cumulatively very low for the vast majority of passengers who
travel infrequently during the year.
Nevertheless, some airlines that fly polar routes, such as United Airlines, are sensitive to
solar storm events, not necessarily for the added radiation load, but for the disruption of
emergency high-frequency communications with ground controller, which violates FAA safety
regulations. For example, on January 20, 2005 a severe solar storm event caused 26 United
Airlines flights to be detoured to lower altitudes and latitudes. The steady increase in the
number of polar routes suggests a larger number of people will be affected by such events as
time goes on. As Figure 7 indicates, currently, 1.7 million passengers travel these routes each
year [Murtagh, 2010].
The most recent, well publicized event where airlines were diverted to other routes came
with the powerful January 23, 2012 solar flare and CME. Hailed as the biggest solar storm
since 2003, Delta Airlines chose to divert its flights to more southerly routes, while American
Airlines took no operational action [Waugh, 2012]. Despite the impacts to airline
communications and flight safety, there are no instances where space weather has affected
passengers or airline flight crews, so in some sense, the issue of flight safety and space
weather presents a very modest health risk, but ironically a significant cost in flight time and fuel
to airline companies that choose to apply mitigation.
-----------------------------------------
Insert Figure 7 about here
-------------------------------------------
2.0 Forecasting Strategies
The specific components of the space weather environment that are known to cause
human impacts are solar x-ray flares, coronal mass ejections, solar proton events, geomagnetic
storms, galactic cosmic rays, electrostatic discharges, and energetic particles in the
magnetosphere. X-ray flares cause ionospheric changes and upper atmosphere heating,
which cause problems for LEO satellites and GPS-based systems. During the impulsive phase
of a CME, shocks also form that lead to the acceleration of particles and solar proton events.
Some of these energetic particles can also arise from the site of the solar flare itself. Solar
proton events and galactic cosmic rays can have energies of 10s of MeV, which lead to SEUs
in computer circuitry, or to enhanced radiation exposure by airline passengers and astronauts (.
The complexity of the space weather environment, and the many ways in which it can
invade our technology to cause problems, almost precludes that we will ever be able to start
from an initial set of solar or geophysical data and use this to determine whether a specific
transformer or satellite system will fail. Consider that a satellite can be rendered inoperable if a
single energetic particle causes a permanent failure of a critical gate, and the vast number of
these particles that pass through a satellite’s volume during its operational lifetime. In quantum
electrodynamics, arguably the most precise physical theory known, precisions of 1 part per 10
billion are routine, but in terms of that fatal energetic particle, we would need a predictive
algorithm of nearly the same caliber, otherwise we are forced to always deal in probabilities.
We also know that, in the space sector with over 2000 operating satellites, it is a rare
event for satellites to actually fail during solar storms. How is it that, in the most recent Galaxy-
15 or Anik-F6 outages, other satellites nearby were not similarly disturbed? We see this curious
paradox again and again in reports of satellite outages related to space weather events. The
industrial response is that satellite failures have much less to do with external space weather
events, which reasonably should have affected more than one satellite simultaneously, than
with manufacturing or software problems internal to the satellite itself such as tin wisker growth
(e.x. Galaxy-7 ca 2000), solar panels designs (e.x. Tempo-2 ca 1997), or software errors (e.x.
Galaxy-15 ca 2010). This raises another important issue that has been a much-discussed topic
among space weather forecasters. How can you predict which space weather events will be
important if the various industries that control the vulnerable assets are not transparent with
their anomaly data?
For most types of space weather events, by the time they are detected it is already too
late to mitigate. This is the case for all of the events that flow from solar or cosmic energetic
particles, or solar X-ray flares. Phenomena such as CMEs, on the other hand, provide us with 1
to 3 days notice of arrival near earth once they are spotted leaving the solar vicinity. To keep
forecasting costs low, what we would like to do is to come up with a small number of
inexpensive measurement indices of the solar and geophysical environment, and through some
TBD algorithm, convert these into a statement about whether a particular resource or asset is in
eminent danger and what the nature of that danger might be so that industry can take action.
We also want to minimize the number of false alarms which can be costly and result in
progressive lack of confidence in the forecasts themselves.
2.1 Solar Storms – flares and CMEs
The simplest correlations we can search for involve solar flares, CMEs, SPEs and the
sunspot cycle, because sunspots can be inexpensively counted and studied with ground-based
instrumentation. We are reasonably certain that solar flares and magnetic reconnection events
require concentrated photospheric magnetic fields, which manifest themselves as sunspots, so
we expect more of these events during times of sunspot maximum than sunspot minimum. Yet
even during sunspot minimum the number of significant flare events is not zero as for example
the M6.4 flare on February 7, 2010, or the X2.6 flare on July 9, 1996. So if you are operating a
GPS system or a WAAS system and require 100% coverage for safety, you will need a back up
plan even during sunspot minimum. What about coronal mass ejections and the effect they can
have on electric power grids?
According to an analysis of 314 halo CMEs during the 23rd cycle by Tripathi and Mishra
(2005), about 3 occurred during sunspot minimum (1996) and 61 during sunspot maximum
(2001), yet fewer than 1 in 6 halo CMEs led to significant geomagnetic storms!
Insert Figure 8 about here
Odenwald (2007) created a combined data base of X-class flares, halo CMEs and Solar
Proton Events (SPE) for the period January-1996 to June-2006, during a time in which
SOHO/LASCO detected 11,031 coronal mass ejections. Of these, 1186 were nominally 'halo'
events including back-side ejections, however, only 598 were actually directed towards Earth.
During the same period of time, 95 solar proton events were recorded by the GOES satellite
network orbiting Earth. Of these SPEs, 61 coincided with Halo CME events. Solar flares were
also recorded by the GOES satellites. During this time period, 21,886 flares were detected, of
which 122 were X-class flares. Of the X-class flares, 96 coincided with Halo CMEs, and 22 X-
class flares also coincided with 22 combined SPE+Halo CME events. There were 6 X-flares
associated with SPEs but not associated with Halo CMEs. A total of 28 SPEs were not
associated with either Halo CMEs or with X-class solar flares. The result can be summarized in
the Venn diagram shown in Figure 8.
What this simple statistical exercise shows is that many of X-class flares (20 of 122),
halo CMEs (458 of 598) and SPEs (28 of 95) are maveric events not associated in general with
the other two types of phenomena. One cannot use halo CMEs to predict if an SPE will result
(only (39+22)/598 = 10 % of the time). One cannot use X-class flares to determine whether an
SPE will result (28/122 = 23% of the time), or using halo CMEs to predict X-class flares, (96/598
= 16% of the time), but if an X-class flare is seen, then you have a (96/122 = ) 79% chance that
a halo CME results, or a (28/95 = ) 29% chance that an SPE results. These statistical results
demonstrate that the path to any sensible form of space weather prediction will probably always
be fraught with the shear uniqueness of each and every space weather event, and the way that
it is then ‘transduced’ by a myriad of technological platforms whose properties and
succeptabilities are often out of the public domain. This is far less like the tornado that rumbles
through a state and wreaks havoc, than a lightning storm whose strike points are utterly random
on the landscape and damage or death is literally a matter of bad luck. This may well be the
situation for global forecasting, but at the individual active region-scale, the situation is
fortunately much more optimistic.
2.2 Reliability of X-class forecasts.
Solar flares and CMEs are the most dramatic precursors of transient changes in space
weather conditions, and considerable effort has been expended in developing predictive
schemes for them (for a review see Forbes, 2010). In all cases, the ability to predict whether a
flare or a CME will occur depends on the quality of the data gathered through the deployment of
sophisticated, and expensive equipment. CMEs cannot be studied without space-based
chronographic equipment (e.g. SOHO/LASCO), sensitive photometers that detect scattered light
from them in transit (STEREO), or in-situ particle and field measurements made at L1 ( e.g.
ACE). There are no ground-based techniques for studying CMEs that could lead to significant
cost savings over using space-based assets that need to be replaced every 10 years or so.
Similarly for solar flares and the solar x-ray emission, only space-based sensors provide the
data required to detect and quantify their severity, with no ground-based analogues to the x-ray
technology. The good news is that both CMEs and x-ray flares produce distinctive ‘Type-II’
bursts of radio-wavelength radiation that ground-based radio telescopes can profitably detect
(e.g. Gopalswamy et al., 2005). By combining ground-based and space-based data over the last
20 years, considerable progress has been made in developing reliable forecasting algorithms
which can provide nearly 100% certainty over the next 24-hour period for significant CME or
flare activity. Most rely on a description of the topology and morphology of sunspots and their
pre-cursor magnetic fields.
Solar flares have been studied extensively since the 1930s since they are historically
known to cause shortwave outages. It has been understood for some time that sunspots with
complex field topologies are prime candidates for flaring activity (Hudson, 2010). Modern-day
analyses that incorporate precursor information about changes in sunspot field topology such
as rotation and shear, and past time history of activity (Nunez et al, 2005), the McIntosh
classification of the sunspot group (Gallagher et al., 2002) lead to forecasts of X-class flares in
the next 24-hour period that are more than 90% reliable with few false-positives. More recently,
Colak and Oahwaji (2007) use a neural network approach to achieve prediction accuracies of
92% for occurrence and 88% for classification (M or X-class). None of these statistical methods
actually employ any physics-based knowledge of the underlying flaring process, but merely
search for correlations among a diverse ensemble of parameters available in various data
bases and archives.
Detailed measurements of the 3-d shape of active region, surface magnetic fields and
their classification (e.g the Wilson classifications), has led to most of the advances in flare
forecasting during the previous sunspot cycle. The legacy of this surface-field approach is best
shown in the research by Steward et al. (2011) who used data from the National Solar
Observatory’s, Global Oscillation Network Group (GONG) to investigate strong-gradient polarity
inversion lines, and neutral lines in maps of solar magnetic fields near active regions. By
classifying each active region seen in 2003 according to a 5-parameter scheme (e.g., field
gradient strength, curvature, neutral line length), they statistically compared 44 combinations of
these 5 indices and found an optimal set that maximized the accuracy of predicting a flaring
event. The resulting, optimized algorithm, called FlareCast, can predict a flare with 88%
confidence within a 24-hour period, with a 10% false-positive rate (i.e. 1 prediction out of 10 will
turn out not to occur).
But events transpire so rapidly that, by the time surface fields start to become twisted
into the pre-flare state, it is already too late to prepare Earth-based systems for the resulting
burst of X-rays and energetic particles, which arrive in less than an hour. During the first decade
of the 21st century, a steadily accumulating archive of sub-surface active region images from
the GONG program, has allowed changes in the plasma flows some 65,000 km below the
surface to be studied at a 10-minute cadence, for over 1,000 active regions. In a path-breaking
paper by Reinard et al. (2010), reoccurring sequences of magnetic topology change were
discovered that presaged surface flaring events, with a lead time of 2 to 3 days before the
surface eruption occurred. A vorticity-based index allows active regions to be classified in terms
of its future activity, and can discriminate between regions producing C, M and X-class flares.
The study supports the idea that rotational kinetic energy twists sub-surface fields into unstable
configurations, which are then involved in explosive magnetic reconnection at the surface.
Coronal mass ejections have a shorter history in the space weather community since
their role in the dynamics of the geomagnetic field was only deduced in the 1980s, and actual
space weather effects were not observed until the Quebec Blackout in 1989. X-class flares
have been studied as precursors for CMEs (HaiminWang et al., 2002) as have sigmoidal
features revealed in soft X-ray imaging of the corona (Sterling, 2000; Canfield et al., 1999), and
vector magnetogram studies of the length of the ‘strong-field, strong-shear’ main neutral line in
sunspot groups (Falconer, 2001). Actual predictive schemes remain lacking due to the failure to
bridge the gap between the changes in the small-scale surface fields near measurable active
regions, and the often hidden large-scale magnetic field rearrangements that occur before the
CME is launched.
Once evidence is available for a CME launch, the arrival time at Earth and the
geoeffectivness become important predictive issues. Some progress has been made in this
area. It has been known for several decades that the most efficient energy transfer into the
geomagnetic field occurs for ‘south directed’ CME field orientations. This can currently only be
determined by in situ measurements made by satellites at L1 such as ACE. Also, the speed of
the CME and the quantity of entrained plasma determine the ram pressure of the arriving CME
at the magnetosphere boundary. The effect of aerodynamic drag by the interplanetary medium
has also been studied (Song, 2010), and the results provide significantly improved
determinations of the initial CME speed, its speed at 1 AU, and the transit time. The advent of
STEREO spacecraft imaging of CMEs at large angles from the sun-earth axis have verified the
deceleration of fast-moving CMEs in the interplanetary medium, and that CMEs need to be
tracked at least 30 degrees from the sun in order to obtain arrival time accuracies less than
about 6 hours.
Solar proton events, often associated with shock acceleration in the initial stages of CME
ejection, have also entered the domain of forecasting through the same statistical studies that
proved successful with X-ray forecasting. For instance, Chin (2005) used an archive of 28 SPE
between 1997-2000 and compared these events with solar radio bursts recorded between 245
MHz and 15,400 MHz to find a strong correlation between Type-III radio bursts at 245 MHz and
the appearance of an SPE observed some 1-2 days later.
2.3 The current solar cycle – and beyond.
It is an elementary exercise to follow the general trending of the last 23 sunspot cycles
and predict that the current one will have a roughly-11 year period, followed no doubt by, Cycle
25, 26, etc into the future. Of course we have historical evidence from the Maunder Minimum
that, on occasion, the 11-year cycle seems to vanish entirely. A diminished solar activity cycle
means that there are proportionately fewer CMEs, flares and SPEs to worry about, and so one
expects a significantly lower risk of GICs in the electric power grid, satellite anomalies and a
lowered risk for excessive radiation exposure for astronauts. Since the flux of cosmic rays is
anti-correlated with the solar activity cycle, this still means that we will experience enhanced
cosmic rays fluxes.
In 2006, a number of prediction schemes were exercised in order to forecast what Cycle
24 might be like, but it was soon recognized from the steady pile up of ‘spotless days’ that Cycle
24 would be qualitatively different than any cycle since ca 1906. There was even some
speculation that a new Maunder Minimum might be at hand. A survey by Pesnell (2008) shows
that there were over 50 predictions for the peak year and SSN for Cycle 24 using many
different, and often independent approaches, both statistics-based and physics-based (see
Charbonneau, 2010). The earliest ‘physics –based’ predictions have suggested that Cycle 24
might be as severe as Cycle 23, and perhaps even worse than any cycle in the last 400 years.
This turned out not to be in line with the trends that began to materialize by ca 2010 and now
estimates have been revised downwards to a peak near 60 and a peak time during the second-
half of 2013.
Recent studies using the GONG network had not detected the sun’s torsional oscillation,
which usually is an advanced indicator of east-west, subsurface plasma flows, and that have
been identified as the harbingers of sunspot formation in the next cycle. Before Cycle 24 had
begun, these same torsional oscillations had been detected and correctly predicted the late-
arrival of the onset of Cycle 24. The concern is that the start of Cycle 25 may be significantly
delayed to ca 2022, or possibly not occur at all [Penn and Livingston, 2006; Hill, 2011].
3.0 Modeling the Economic and Societal Impacts.
Compared to terrestrial weather, space weather as we have experienced it during the
Space Age, is in the ‘noise’ in terms of economic impact. The difference between 1 or 2
extreme weather events in 2011, out of the 37 recorded world-wide, is about $2 billion. This
equals the economic impact of ALL the commercial satellites lost in the entire 11-year solar
cycle ending 2006!
Cycle 23 will be seen by historians, no doubt, as a watershed moment in space weather
history. Prior to Cycle 23, there was little or no public discussion about space weather
vulnerability during the Space Age, although our grandparents surely knew all about the
practical consequences of space weather and the insufferable short wave outages. With Cycle
23, we had SOHO providing the public with dazzling and ominous movies of solar storms, and
many popularizers, including myself, who went on the stump to sort out for the public all the
ways in which we could be affected. Then, just before the famous Halloween Storm of 2003, we
had the first high-profile Congressional hearing about space weather in the context of why
NOAAs Space Environment Center (SEC) budget should not be halved. Once Homeland
Security became involved, we then had a new round of hearings about our infrastructure
vulnerability to space weather events. The Space Weather Forum was held in Washington, DC
on Capitol Hill in June 2008 to educate Capitol Hill about space weather issues. Meanwhile,
researchers began the difficult task of trying to quantify what these impacts could cost us and
the social disruption that might follow.
Kappenmann (1997) has an extensive record of modeling the US power grid with increasingly
more sophisticated models of the electrodynamics of GICs and exhaustive studies of the North
American electric grid network at the component level. Currently, his efforts use historical
geomagnetic storms (e.g. 1921 event) and their impact on the contemporary electric power grid.
Among the forecasts are for year-long recovery periods costing over $1 trillion in GDP.
Teisberg and Weiher (2000) estimated that the economic benefits of providing reliable warnings
of geomagnetic storms to the electric power industry (alone) would be approximately $450
million over three years (note that this doesn't include any other impacted industries). This is
well above the $100 million cost of a new operational satellite that would provide such warnings
(ACE, Triana)
Odenwald and Green (2007) modeled the economic losses to commercial satellites in LEO,
MEO and GEO orbits and deduced that an 1859-scale ‘superstorm’ arriving near sunspot
maximum could cost $50 billion in lost revenue and assets.
August 1988, Oak Ridge National Laboratory and the NRC published ‘Evaluation of the
Reliability for the Offsite Power Supply as a Contributor to the Risk of Nuclear Plants’. This set
the stage for considering the impact of space weather-related GICs on the reliability and safety
of nuclear power plants (Kirby et al., 1988).
April 1989, Northwest Power Coordinating Council (NPCC) approved the document
"Procedures for Solar Magnetic Disturbances Which Affect Electric Power Systems" which has
been updated several time.(NPCC, 1989)
October 2003 – ‘What is Space weather and who should forecast it? Congressional Hearing on
Space Weather held before the Subcommittee on Environment, Technology, and Standards,
Committee on Science, House of Representatives, One Hundred Eighth Congress, first session,
October 30, 2003, (Congress, 2003)
December 2005, Idaho National Laboratory and NRC published ‘Reevaluation of Station
Blackout Risk at Nuclear Power Plants--Analysis of Station Blackout Risk.’ The executive
summary from this report reads in part: The availability of alternating current (ac) power is
essential for safe operations and accident recovery at commercial nuclear power plants. (INL,
2005)
April, 2008: "Report of the Commission to Assess the Threat to the United States from
Electromagnetic Pulse (EMP) Attack: Critical Infrastructures". The US Congress funded a
vulnerability assessment research under the National Defense Authorization Act to evaluate the
impact of an electromagnetic pulse (EMP) from a high altitude nuclear detonation by a terrorist
event on the nation's critical infrastructure including the electric grid. The same study also
discussed geomagnetically-induced currents. (EMP Commission, 2008)
2008 ‘Severe Space Weather Events—Understanding Societal and Economic Impacts
Workshop Report’. The National Academy of Sciences determined that severe geomagnetic
storms have the potential to cause long-duration outages to widespread areas of the North
American grid. (NAS, 2008)
June 2010, "High-Impact, Low-Frequency Event Risk to the North American Bulk Power
System," jointly sponsored by NERC and the Department of Energy, NERC now concedes that
the North American power grids have significant reliability issues in regard to High-Impact, Low-
Frequency events such as severe space weather. The NERC report explains commercial grid
vulnerability to space weather (NERC, 2010)
October 2010, ‘Electromagnetic Pulse: Effects on the U.S. Power Grid’, Oak Ridge National
Laboratory released a series of comprehensive technical reports for the Federal Energy
Regulatory Commission (FERC) in joint sponsorship with the Department of Energy and the
Department of Homeland Security. These reports disclose that the commercial power grids in
two large areas of the continental United States are vulnerable to severe space weather. The
reports conclude that solar activity and resulting large earthbound CME, occurring on average
once every one hundred years, would induce a geomagnetic disturbance and cause probable
collapse of the commercial grid in these vulnerable areas. The replacement lead time for extra
high voltage transformers is approximately 1-2 years. As a result, about two-thirds of nuclear
power plants and their associated spent fuel pools would likely be without commercial grid
power for a period of 1-2 years. (Oak Ridge Labs, 2010)
Armed with all this bad news, and with the storms of Cycle 24 now beginning, it has
become commonplace for Reporters to quote these studies and offer titles such as ‘A big solar
storm could cost $2 trillion, could be a global Katrina’ or ‘Solar storm buffets Earth: How
protected is the US power grid?’. The danger is that, through constant repetition of this
Doomsday theme, the public will become inured to the message in the face of the inevitable
false alarms such as the January 2012 storm. While it is certainly important to keep the
preparation message alive given the consequences to our infrastructure, as scientists and
space weather forecasters, we need to be more careful with delivering this complex message to
a Public increasingly eager for a binary answer to their safety.
References:
Amin, Massoud: "Powering the 21st Century: We can -- and must -- modernize the grid"; Today's Engineer: Mar 2005. http://www.todaysengineer.org/2005/Mar/grid.asp
Patterson, T., 2010, ‘US Electricity Blackouts Skyrocketing’, October 15, 2010, http://www.cnn.com/2010/TECH/innovation/08/09/smart.grid/index.html, accessed April 20, 2012.
Baker ,D. 2000 "The occurrence of operational anomalies in spacecraft and their relationship to space weather", IEEE Transactions on Plasma Science, v. 28, p. 6.
Bedingfield ,K. L., Leach, R.D., and Alexander, M.B. ,1996, "Spacecraft System Failures and Anomalies Attributed to the Natural Space Environment", NASA Reference Publication 1390, August 1996.
Belov, A., L. Dorman, N. Iucci ,O. Kryakunova ,and N. Ptitsyna, 2004, 'The relation of high- and low-orbit satellite anomalies to different geophysical parameters', in 'Effects of Space Weather on Technology Infrastructure', Ioannis A. Daglis ,ed., Kluwer Academic Publishers, pp.147-163.
Berg, Jerome, 2008, ‘Broadcasting On The Short Waves, 1945 To Today’, (McFarland, NY), p. 22.
Berg Insight, 2009, '295 million GPS Handsets Sold in 2010; 940 million Expected in 2015 ', April 19, 2011, http://www.gpsbusinessnews.com/tags/Berg%20Insight/, accessed April 20, 2012.
Berkowitz, B., 2011, ‘Extreme weather batters the insurance industry’, February 09, 2011, http://www.reuters.com/article/2011/02/09/us-insurance-climate-idUSTRE7182XG20110209?pageNumber=1, accessed May 4, 2012.
Boston.com, 2012, ‘Computer glitch hits Brazil’s biggest airline’, March 02, 2012, http://articles.boston.com/2012-03-02/news/31117691_1_computer-glitch-check-in-system-airports, Accessed April 15, 2012.
Bottollier-Depois JF, Chau Q, Bouisset P, Kerlau G, Plawinski L, Lebaron-Jacobs L., 2000, ‘Assessing exposure to cosmic radiation during long-haul flights.’, Radiation Research, v. 153, pp. 526-532.
Brekke, P., M. Chaloupy, B. Fleck, S.V. Haugan, T. van Overbeek, and H. Schweitzer, 2004, “Space weather effects on SOHO and its space warning capabilities”, in Effects of space weather on technology infrastructure’, I. A. Daglis (ed), Kluwer Academic Publishers, pp 109-122.
Canfield, R.C., Hudson, H.S. and McKenzie, D.E.,1999 ‘Sigmoidal morphology and eruptive solar activity’ Geophysical Reserach Letters, 26 (1999), pp. 627–630
Careless, James, 2011, ‘Changes Continue for Shortwave’, 10.26.2011, http://www.rwonline.com/article/changes-continue-for-shortwave/24684, accessed on April 15, 2012.
Careless, James, 2010, ‘Whatever Happened to Shortwave Radio? ‘, 03.08.10, http://www.rwonline.com/article/whatever-happened-to-shortwave-radio/2842, Accessed on April 19, 2012.
CBS News ,2012, ' Satellite problems ground Nunavut flights', October 6, 2011, http://www.cbc.ca/news/technology/story/2011/10/06/north-satellite-phone-outage.html?cmp=rss, accessed April 15, 2012.
Charbonneau, P., 2010, 'Modeling solar and stellar dynamos', in 'Heliophysics: Evolving solar activity and the climates of space and earth', Carolus Schrijver and George Siscoe, eds., (Cambridge University Press; United Kingdom), pp.141-177.
Chin, J., 2005, ‘Evidence for a Strong Correlation of Solar Proton Events with Solar Radio Bursts’, J. of Astron. Astrophys. V. 5, p.110.
Cho, M. and Y. Nozaki, 2005, “Number of arcs estimated on solar array of a geostationary satellite”, J. Spacecr. Rockets. V. 42, N. 4, pp. 740-748.
Clark, S, 2010, ‘Galaxy 15 'zombiesat' still alive after expected off date ‘,September 15, 2010, http://www.spaceflightnow.com/news/n1009/15galaxy15/, accessed April 21, 2012.
Colak, T. and Qahwaji, R. ,2007, "Automated Prediction of Solar Flares Using Neural Networks and Sunspots Associations," Advances in Soft Computing, DOI 10.1007/978-3-540-70706-6, Springer, 39,pp. 316-324, 2007. http://spaceweather.inf.brad.ac.uk/journals/wsc11_tc_rq.pdf, accessed April 20, 2012.
Congress, 2003, What is space weather and who should forecast it?: hearing before the Subcommittee on Environment, Technology, and Standards, Committee on Science, House of Representatives, One Hundred Eighth Congress, first session, October 30, 2003, Volume 4, http://www.solarstorms.org/CongressSW.html, accessed, April 23, 2012.
Davis, C.J., Kennedy, J. and Davies, J.A., 2010, ‘Assessing the Accuracy of CME Speed and Trajectory Estimates from STEREO Observations Through a Comparison of Independent Methods’, Solar Physics, vol. 263, pp. 209-222.
De Shelding, P. 2011, ‘Electrostatic discharge crippled Galaxy 15 Intelsat says’, January 13, 2011, http://www.spacenews.com/satellite_telecom/110113-electrostatic-discharge-galaxy15.html, accessed April 15, 2012.
Department of Energy: "Gridworks"; Office of Electricity & Energy Reliability: 2005. http://www.energetics.com/gridworks/index.html
Doherty, P.H., 2011, ‘Space Weather Effects on GPS and WAAS’, August 5, 2011, http://www.bc.edu/research/isr/spaceweathereffects.html, accessed April 15, 2012.
Dorman, L. I. et al., 2005, "Space weather and space anomalies", Annales Geophysicae ,V. 23, pp. 3009-3018.
Duncan, G., 2010, ‘Use a cellphone? Join the rest of mankind!’, http://www.digitaltrends.com/mobile/two-thirds-of-all-humans-use-a-cell-phone/, accessed May 2, 2012.
Eaton Corporation, 2011, ‘Blackout Tracker: 2011 Annual Report’, http://powerquality.eaton.com/info/GenOutput.asp?Quest_user_id=657593&leadG_Q_QRequired=False&menu=, April 20, 2012.
Electromagnetic Pulse Commission, 2008, ‘Report of the Commission to Assess the Threat to the United States from Electromagnetic Pulse (EMP) Attack’, April 2008, http://www.empcommission.org/docs/empc_exec_rpt.pdf, accessed April 20, 2012.
Energy Information Agency (EIA): "Electric Power Annual with data for 2005"; EIA/DOE: revised 09.11.2006. http://www.eia.doe.gov/cneaf/electricity/epa/epa_sum.html
Falconer, D.A., 2001, ‘A prospective method for predicting coronal mass ejections from vector magnetograms’, JGR, v. 106, No. A11, pp 25,185 – 25,190.
Farthing, W. H., J.P. Brown and W.C. Bryant (1982: NASA Tech Memo 83908, March 1982)
Fennell, J. F. et al. ,2000 "Internal charging; A preliminary environmental specification for satellites", IEEE Trans. On Plasma Physics ,V. 28, pp. 2029.
Ferris, D. L. ,2001, "Characterization of Operator-Reported Discrepancies in Unmanned On-Orbit Space Systems", Masters Thesis, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, June 2001.
Ferster, W., 2005, 'Satellite industry revenue jumps 6 percent in 2003', Space News, July 7, 2005.
Forbes, T, 2010, 'Models of coronal mass ejections and flares', in 'Heliophysics: Space storms and radiation- causes and effects', Carolus Schrijver and George Siscoe, eds., (Cambridge University Press; United Kingdom), pp.159-191.
Futron ,2002, "Satellite Insurance Rates on the Rise - Market correction or overreaction?", July 10, 2002.
Gallagher, P. T., Moon, Y.-J., & Wang, H. 2002, Sol. Phys., 209, 171.
Gopalswamy, N., Aguilar-Rodriguez,E., Yashiro,S., Nunes,S., Kaiser,M.L. and Howard, R.A., 2005, ‘Type II radio bursts and energetic solar eruptions’, J.G.R., v. 110, A12S07, doi:10.1029/2005JA011158, 2005., http://cdaw.gsfc.nasa.gov/publications/gopal2005-turku.pdf, accessed May 2, 2012.
Hamachi-LaCommare, K. and Eto, J., 2004, ‘Understanding the Cost of Power Interruptions to U.S. Electricity Consumers’, http://certs.lbl.gov/pdf/55718.pdf, accessed April 21, 2012.
HaiminWang ,T.J., Spirock, J.Q., Haisheng, J., Yurchyshyn, V., Moon, Y-J., Denker, C. and Goode, P.R., 2002, ‘Rapid changes of magnetic fields associated with six, X-class flares’, Ap.J., v. 576, pp. 497-504.
Hannah, Eric, 2004, ‘Cosmic ray detectors for integrated circuit chips’, US Patent 7,309,866; http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=7,309,866.PN.&OS=PN/7,309,866&RS=PN/7,309,866. Accessed April 15, 2012.
Hecht, H. and Hecht, M. 1986, "Reliability Prediction for Spacecraft", Rome Air Development Center, Griffiss Air Force Base, New York, RADC-TR-85-229, December, 1985.
Held, Gilbert, 2002, ‘Understanding Data Communications’, (New Riders Publishing, NY), Sixth Edition, Chapter 7.
Hoyos, S.E. ,Evans, H.D.R. and Daly ,E. 2004, "From satellite ion flux data to SEU rate estimation" ,IEEE Trans .Nucl. Sci. v. 51, pp 2927-2935.
Hudson ,H., 2010, 'Observations of solar and stellar flares and jets', in 'Heliophysics: Space storms and radiation- causes and effects', Carolus Schrijver and George Siscoe, eds., (Cambridge University Press; United Kingdom), pp.123-158.
Idaho National Laboratory, 2005, ‘Reevaluation of Station Blackout Risk at Nuclear Power Plants--Analysis of Station Blackout Risk’, December 2005, http://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr6890/, accessed April 20, 2012.
International Cable Protection Committee (ICPC), 2009 ‘Fishing and Submarine Cables Working Together’, http://cil.nus.edu.sg/wp/wp-content/uploads/2009/10/ICPC-Fishing-Booklet-090223.pdf, Accessed on April 18, 2012.
JPL, 2012 'Ionospheric TEC Map', April 20, 2012,http://iono.jpl.nasa.gov/latest_rti_global.html, accessed April 20, 2012.
Kappenman, J., 2012, ‘Impact of Severe Solar Flares, Nuclear EMP and Intentional EMI on Electric Grids’, The Electric Infrastructure Security Summit, Third Annual World Summit, London, ‘http://www.eissummit.com/images/upload/conf/media/EIS_Kappenman_Part1.pdf‘, accessed April 20, 2012.
-------------------, 2010, ‘Low-frequency protection concepts for the electric power grid: Geomagnetically Induced Current and E3 HEMP mitigation’, January, 2010, http://www.ferc.gov/industries/electric/indus-act/reliability/cybersecurity/ferc_meta-r-322.pdf, accessed April 20, 2012.
-------------------, 1997, ‘Geomagnetic storm forecasts and the power industry’, EOS Transactions. January 29, 1997 p. 37.
Karn, R., 2007, ‘Electrifying Change’, April 4, 2007, http://www.energybulletin.net/node/29104, accessed Apriul 15, 2012.
Kirby, B.J., Kueck, J.D. and Poole, A.B., 1988, ‘Evaluation of the Reliability for the Offsite Power Supply as a Contributor to the Risk of Nuclear Plants ‘,ORNL/NRC/LTR-98/12, http://www.consultkirby.com/files/NRC.pdf, accessed April 23, 2012.
Koons, H.C. 2000, "The Impact of the space environment on space systems", Proceedings of the 6th Spacecraft Charging Technology Conference ,Air Force Research Laboratory
Koons, H.C., Mazur ,J.E., Selsnick ,R.S., Blacke, J.B. ,Fennell, J.F. ,Roeder, J.L, and Anderson, P.C. ,1999, "The Impact of the Space Environment on Space Systems", Aerospace technical Report, TR-99(1670)-1, July 20, 1999. Space and Missle Systems Center, Air Force Materials Command ,2430 E. El Segundo Boulevard, Los Angeles Air Force Base, CA, 90245.
Lanzerotti ,L..J., Gary, D.E., Nita, G.M., Thomson, D.J. and Maclennan, C.G, 2005, 'Noise in wireless systems from solar radio bursts', Advances I nSpace Research, Vol. 36, No. 12, pp.2253-2257. http://solar.njit.edu/preprints/gary1218.pdf, accessed April 18, 2012.
Mack, Eric, 2011, 'Major satellite outage affecting ISPs, ATMs and flights', October 6, 2011, http://news.cnet.com/8301-1023_3-20116846-93/major-satellite-outage-affecting-isps-atms-flights/, accessed April 15, 2012.
Marowits, Ross, 2011, ' Satellite outage disrupts communications in Canada’s North, October 6, 2011, http://www.jocosarblog.org/jocosarblog/2011/10/satellite-outage-disrupts-communications-in-canadas-north.html, accessed April 15, 2012.
Martin, Guy, 2012, 'SumbandilaSat beyond repair', January 25, 2012, http://www.defenceweb.co.za/index.php?option=com_content&view=article&id=22870:sumbandilasat-beyond-repair&catid=90:science-a-technology&Itemid=204, Accessed April 20, 2012.
McCracken, K.G., G.A.M. Dreschoff, E.J. Zeller, D.F. Smart, and M.A. Shea (2001), Solar cosmic ray events for the period 1561-1994: 1. Identification in polar ice, 1561-1950, J. Geophys. Res., 106(A10), 21,585-21,598.
Murtagh ,W., 2009, 'Space Weather; The next frontier in Severe Impacts', April 2, 2009, http://www.ametsoc.org/atmospolicy/documents/Bill_Murtagh_SpWx.pdf, accessed April 15, 2012.
Murtagh, B., 2010, ‘Space weather impacts on aviation systems’, September 8, 2010, http://www.easa.europa.eu/conferences/iascc/doc/Workshop%201%20Presentations/Workshop1_DAY%201/3_Murtagh_NOAA/Space%20Weather%20Impacts%20on%20Aviation%20Systems.pdf, accessed April 15, 2012.
National Academy of Sciences, 2008, ‘Severe Space Weather Events—Understanding Societal and Economic Impacts Workshop Report’, http://www.nap.edu/catalog.php?record_id=12507, accessed April 20, 2012.
New York Times, May 16, 1921, ‘Sunspot credited with rail tieup ‘, p.2.
-----, November 2, 1903, ‘…’, p.2.
-----, July 8, 1941, ‘Shortwave channels to Europe are affected’, p. 10.
-----, September 19, 1941, ‘ Major baseball game disrupted’, p. 25.
-----, July 19, 1947, ‘ Sunspots delay planes’, p.15.
-----, February 21, 1950, ‘Sun storm disrupts radio cable service’, p. 5.
-----, August 20, 1950, ‘Radio messages about Korean War interrupted’, p. 5.
-----, April 18, 1957, ‘World radio signals fade’, p. 25.
-----, February 11, 1958, ‘Radio blackout cuts US off from rest of world’, p. 62.
NERC, 2010, ‘High-Impact, Low-Frequency Event Risk to the North American Bulk Power System’, June 2010, http://www.nerc.com/files/HILF.pdf, accessed April 20, 2012.
NGDC, 2007, http://goes.ngdc.noaa.gov/data/plots/
NGDC, 2007, Ap-Kp data from http://www.ngdc.noaa.gov/stp/GEOMAG/kp_ap.html
NOAA, 2001, ' Severe Space Weather Storms Affect GPS Meteorology Network', April 23, 2001, http://www.esrl.noaa.gov/gsd/media/hotitems/2001/01Apr23.html, accessed April 15, 2012.
NOAA, 2004: Intense space weather storms: October 19 – November 07, 2003. Service Assessment, 60pp. http://www.nws.noaa.gov/os/assessments/pdfs/SWstorms_assessment.pdf
NOAA Magazine, 2007, ' Researchers find global positioning system is significantly impacted by powerful solar radio burst’, April 4, 2007, http://www.noaanews.noaa.gov/stories2007/s2831.htm, accessed April 15, 2012.
NPCC, 1989, ‘Procedures for Solar Magnetic Disturbances Which Affect Electric Power Systems’, C-15, https://www.npcc.org/Standards/Procedures/Forms/Public%20List.aspx, accessed April 20, 2012.
Nunez, M., Fidalgo, R., Baena, M. and Morales, R., 2005, ‘The influence of active region information on the prediction of solar flares: an empirical model using data mining’, Annales Geophysicae, v. 23., pp 3129-3138., http://www.ann-geophys.net/23/3129/2005/angeo-23-3129-2005.pdf, accessed April 22, 2012.
Oak Ridge Laboratory, 2010, ‘‘Electromagnetic Pulse: Effects on the U.S. Power Grid’, October 2010, http://www.ferc.gov/industries/electric/indus-act/reliability/cybersecurity/ferc_executive_summary.pdf, accessed April 20, 2012.
Odenwald, S.F., J. Green, W. Taylor, ‘Forecasting the Impact of an 1859-calibre Superstorm on Satellite Resources: Part I’, Adv. Sp. Res., v. 38, pp. 280-297.
Odenwald, S. F., and J. L. Green, 2007, Forecasting the impact of an 1859-caliber superstorm on geosynchronous Earth-orbiting satellites: Transponder resources, Space Weather, 5, S06002, doi:10.1029/2006SW000262.
Odenwald, S., 2010, 'Introduction to Space Storms and Radiation', in 'Heliophysics: Space storms and radiation- causes and effects', Carolus Schrijver and George Siscoe, eds., (Cambridge University Press; United Kingdom), pp.15-41.
Odenwald, S. F., 1999, ‘The 23rd Cycle- Learning to live with an active star’, (Westfield Press, Boulder).
OOSR, 1978 "On-Orbit Spacecraft Reliability', PRC Report R-1863, September 30, 1978 and "Analysis of Spacecraft On-Orbit Anomalies and Lifetimes", PRC report R3579, February 10, 1983.
Parnell, Brid-Aine, ‘Epic net outage in Africa as four undersea cables chopped’, 02.28.2012, http://www.theregister.co.uk/2012/02/28/east_africa_undersea_cables_cut/. Accessed on April 18, 2012.
Paxman, Lauren and Dan Hyde, 2011, ‘HSBC systems crash affects millions across UK’, Nov. 5, 2011, http://www.dailymail.co.uk/news/article-2057657/HSBC-customers-outraged-online-b Accessed April 15, 2012.
Penn, M.J. and Livingston, W., 2006, ‘ Long term evolution of sunspot magnetic fields’ IAU 273.
Pesnell, W. D., 2008, ‘Predictions of Solar Cycle 24’, Solar Physics, v. 252, pp 209-220.
Pulkkinen, A Lindahl, S, Viljanen, A and Porjola, R, 2005, ‘Geomagnetic storm of 29-31 October 2003: Geomagneticaly induced currents and their relation to problems in the Swedish high-voltage power transmission system’, Space Weather Vol 3, S08C03 2005
Radio Netherlands, 2011 ‘Solar flare disrupts RNW short wave reception’, 08.09.2011, http://www.rnw.nl/english/bulletin/solar-flare-disrupts-rnw-short-wave-reception. Acessed, April 10, 2012.
Reinard, A., Henthorn, J., Komm, R. and Hill, F., 2010, ‘Evidence that temporal changes in solar subsurface helicity precede active region flaring’, Ap.J. Letters, v. 710, L121-125.
Robbrecht, E., Berghmans, D. and van der Linden, R.A.M., 2009, ’Automated LASCO CME Catalog for Solar Cycle 23: Are CMEs scale invariant?
Robertson, B. and E. Stoneking, 2003, "Satellite GN and C Anomaly Trends", AAS Guidance and Control Conference; 5-9 Feb. 2003; Breckenridge, CO; United States, Report Number AAS-03-071.
Santarini, M., 2005, ‘Cosmic radiation comes to ASIC and SOC design’, May 12, 2005, http://www.edn.com/article/470298-Cosmic_radiation_comes_to_ASIC_and_SOC_design.phpAccessed, April 15, 2012.
Shortwave America, 2012, ‘M9-class Long-Duration Solar X-ray Flare’, 01.23.2012, http://shortwaveamerica.blogspot.com/2012/01/m9-class-long-duration-solar-x-ray.html, accessed on April 12, 2012.
Song, W.B., 2010, ‘An Analytical Model to Predict the Arrival Time of Interplanetary CMEs’, Solar Physics, vol. 261, pp. 311-320.
Space Weather Canada, 2011, ‘Geomagnetic Effects on Communications Cables’, http://www.spaceweather.gc.ca/se-cab-eng.php,. Accessed April 18, 2012.
Stassinopoulos, E.G., Brucker, G.J. ,Adolphsen, J.N. and Barth ,J., 1996, "Radiation-Induced Anomalies in Satellites", J. Spacecraft and Rockets, V. 33 ,PP. 877-882.
Sterling, A.C., 2000, ‘Sigmoid CME Source Regions in the Sun: Some recent results’, J. Atmosp. And Sol. Terr. Phys., v. 62, No. 16, pp. 1427-1435.
Steward, G. A., V. V. Lobzin, P. J. Wilkinson, I. H. Cairns, and P. A. Robinson. 2011. Automatic recognition of complex magnetic regions on the Sun in GONG magnetogram images and prediction of flares: Techniques for the flare warning program Flarecast, Space Weather, 9, S11004, doi:10.1029/2011SW000703.
Telegeography, 2012, ‘Submarine Cable Map’, http://www.submarinecablemap.com/, Accessed May 4, 2012.
Tezzaron Semiconductors, 2003, ‘Soft Errors in Electronic Memory’, http://www.tezzaron.com/about/papers/soft_errors_1_1_secure.pdf. Accessed, April 15, 2012.
The Economist, 2009, 'Cutting the cord', August 13, 2009, http://www.economist.com/node/14214847, accessed April 20, 2012.
Townsend, L.W. (2003), Carrington flare of 1859 as a prototypical worst-case solar energetic particle event, IEEE Trans. Nucl. Sci. 50, 2307-2309.
Tribble, A., 2010, 'Energetic particles and technology', in 'Heliophysics: Space storms and radiation- causes and effects', Carolus Schrijver and George Siscoe, eds., (Cambridge University Press; United Kingdom), pp.381-399.
Tripathi, R. and Mishra, A.P., 2005, ‘Properties of halo CMEs in relation to large CMes’, 29th International Cosmic Ray Conference, Prune, http://dpnc.unige.ch/ams/ams_beta/ICRC/ICRC-05/PAPERS/SH14/ind-tripathi-R-abs1-sh14-poster.pdf, accessed April 15, 2012.
Vampola, A. L., 1987, "The aerospace environment of high altitudes and its implications for spacecraft charging and communications". J. Electrostatics ,v. 20, p. 21.
Wahlund, J-E., L.J. Wedlin, T. Carrozi ,A.I. Eriksson ,B. Holback, L. Andersson and H. Laakso, 1999 "Analysis of Freja Charging Events: Statistical Occurrence of Charging Events, WP 130 technical Note (SPEE-WP130-TN) ,February 22, 1999.
Waugh, R., 2012, ‘Biggest solar storm since 2003 pummels atmosphere, forcing planes to divert from northern routes ‘, January 25, 2012, http://www.dailymail.co.uk/sciencetech/article-2091586/Solar-radiation-storm-Flights-diverted-Earths-atmosphere pummeled.html#ixzz1rjThzvhb, accessed April 15, 2012.
Wilkinson, D. 1994, "NOAA's Spacecraft anomaly data base and examples of solar activity affecting spacecraft", Jour. Spacecraft and Rockets, vol. 31, pp 160-165.
Wilkinson, D.C. , Allen, J. 1997, “Satellite Anomaly Data Base”, ftp://ftp.ngdc.noaa.gov/STP/ANOMALIES/
Wik, M., Pirjola ,R., Lundstedt, H. ,Viljanen, A., Wintoft, P. And Pulkkinen, A., 2009, ' Space weather events in July 1982 and October 2003 and the effects of geomagnetically induced currents on Swedish technical systems', Annals of Geophysics, V. 27, pp 1775-1787.
Wrenn, G. L., Rodgers, D.J. and Ryden, K.A., 2002, "A solar cycle of spacecraft anomalies due to internal charging", Annales Geophysicae, v. 20, pp. 953-956.
Xinhuanet, 2005, ‘Solar storm interrupts China's short-wave radio transmission’, 01.21.2005, http://news.xinhuanet.com/english/2005-01/21/content_2491819.htm, Accessed on April 12, 2012.
Xihuanet, 2011, ‘Solar flare affects shortwave radio communications in southern China’, 02.16.2011, http://news.xinhuanet.com/english2010/china/2011-02/16/c_13733621.htm, accessed on April 12, 2012.
Figures and Figure Captions
Figure 1 – The number of short wave stations (vertical axis) has dropped dramatically since the advent of the World Wide Web and other wireless media, which now provide the main source of news reporting in the 21st century. (Data courtesy Careless, 2010)
0
200
400
600
800
1000
1200
1973 1985 1997 2003 2007
S.America
Asia
Africa
Figure 2 – A small portion of a map of the current locations of submarine fiber optic cables ca 2011. (Courtesy TeleGeography,2012)
Figure 3 – This figure shows the effects of high-energy solar protons on an exposed imager on the Solar and Heliospheric Observatory during the January 23, 2012 solar storm. A greatly reduced flux of particles entering shielded satellite circuitry results in SEUs, many of which are harmless, but a few per year can result in serious operational anomalies.
Figure 4 - A new report from the International Telecommunications Union finds that at the end of 2009, 67 percent of all people on Earth were cell phone subscribers (solid line). The number of land line subscribers is now in decline (dotted line) having reached a maximum of 19% of world inhabitants in 2005 (Duncan, 2010).
0
10
20
30
40
50
60
70
80
1998 2000 2002 2004 2006 2008 2010
Year
Pe
r 1
00
pe
op
le
Figure 5 – Total electron content (TEC) calculation for April 20, 2012 for 19:00 UT developed by JPL.
Figure 6‐ GICs cause damage to large transformers through magnetostriction and the resulting thermal heating. This picture is reminiscent of the images of transformers that failed during the 1989 Quebec Blackout (Courtesy Murtagh, 2009)
Figure 7 – The number of passengers flying ‘polar routes’ continues to sharply increase each year to a current level of nearly 1.7 million passengers each year [Murtagh, 2010].
0
1000
2000
3000
4000
5000
6000
7000
8000P
ola
r fl
igh
ts
2000 2001 2002 2003 2004 2005 2006 2007 2008
Year
Figure 8 – A simple Venn diagram showing the frequency of Halo CMEs, X-class flares and Solar Proton Events during Cycle 23. There were a total of 598 Earth-directed CMEs, 95 Solar Proton Events, and 122 X-class flares. It is clear from the intersection statistics that the vast majority of CMEs do not result in SPEs, or are associated with X-class flare events. However, it is also true that the majority of X-class flares are associated with CMEs, and that the majority of SPEs are associated with CMEs as well.